▶ 0:19:32Good morning everyone. Uh I now call the committee on small business to order. Without objection, the chair is authorized to declare a recess of the committee at any time. I now recognize myself for my opening statement. Uh welcome to today's hearing, AI on Main Street. How AI is shaping the future of small business. I want to thank our witnesses uh for coming with us today and and being part of this.
▶ 0:19:56Uh artificial intelligence has quickly transitioned from month from a novelty to an indispensable business tool for main street businesses across the country. AI is streamlining administrative tasks that once required significant time and resources and it's allowing small businesses to focus on growth instead of paperwork. Small businesses are using AI to uh respond quickly to customers, manage inventory, strengthen cyber security, and automate routine paperwork.
▶ 0:20:22As a small business owner myself, I have seen the difference these tools can make, and I know firsthand how much of that burden falls on the owner. Artificial intelligence is no longer something reserved for big tech companies and Fortune 500 boardrooms. Today nearly threequarters of small businesses are using AI in some form and this is uh more than triple the share from just three years ago. So small businesses are closing in on technology gap with the big corporations faster than any prior wave of innovation.
▶ 0:20:52Now importantly early data is showing that AI is rise is not coming at the expense of workers. In fact 82% of small businesses using AI grew their workforce over the past year. Of course, real barriers to broader AI adoption remain. Many owners tell us that they simply don't know how or where to start and that they worry about cost so that they are unsure how to keep their data secure. But one of the biggest barriers in in Washington is is created is called uncertainty.
▶ 0:21:20And the Biden administration took a heavy-handed approach to AI. In October 30, 2023, they released the executive order imposing sweeping regulatory requirements and compliance burdens that that threatened to smother AI innovation before small businesses ever got the chance to benefit from it. We only need to look across the Atlantic to see where that road leads. Europe chose to recogni regulate first and innovate later. And today the small businesses are falling behind and the United States cannot make the same mistake.
▶ 0:21:50Thankfully, the Trump administration reversed course. They resented the Biden approach and replaced it with a pro-inovation strategy through executive order 14179 in America's AI action plan uh clearing away for red doing away with red tape so that AI tools can be adopted with confidence. The SBA is doing its part by also expanding access to AI through a its new Main Street AI accelerator program.
▶ 0:22:16And of course, the committee has led the way with bipartisan legislation such as the AI AI wise act, the AI for Main Street Act, and the SBA artificial intelligence utilization act of 2026. So together, these bipartisan bills ensure that entrepreneurs have access to practical AI and the barriers to AI adoption are reduced. Our goal is to make every small business and entrepreneur make sure they have the opportunity to succeed in the emerging age of AI.
▶ 0:22:45So when small businesses have the tools, they have the training and freedom to innovate, they will out compete, they will out build and out hire anyone in the world and our economy will be stronger because of it. I want to thank our witnesses again for joining us today. I look forward to the conversation ahead and I now yield to ranking member Congressman Cisneros from the great state of California.
▶ 0:23:06Good morning and thank you Mr. Chairman for convening this important hearing. Artificial intelligence will revolutionize the way businesses work and human and humans operate. AI could allow main street businesses to scale their capabilities and perform routine tasks faster and cheaper. However, AI's many risks have incured costs for small firms. AI can produce false outputs, a phenomenon known as hallucinations. further bias and discrimination.
▶ 0:23:34AI decision-making undermines its effectiveness and duly hurts underserved communities and exposes users to legal and reputational harm. And AI companies often take and use the data and information users input into their models for further training and sale to other parties posing privacy risk.
▶ 0:23:54It is challenging to know exactly how to assess and address these problems in part because AI developers through automated machine learning and and most AI developers do not disclose their models training information. I support the consideration and use of AI by our small businesses, but we need to protect our businesses from flexible Small businesses need to be ready to effectively eva evaluate and adopt these tools.
▶ 0:24:21Uneven access to AI and knowledge about how to use it and avoid its risk threatened to cut whole segments of our population off from the benefits of AI. Underserved people have especially fallen behind in their understanding of AI tools, increasing the digital divide and hurting their businesses. It is this committee's responsibility to address these issues. The SBA must provide AI literacy resources to educate entrepreneurs on the risk and potential benefits of AI.
▶ 0:24:50This will reduce the digital divide and equip entrepreneurs with the tools and information to make well-informed, thoughtful decisions about whether AI is right for them, what AI AI can and cannot do, and which specific AI tools are suitable for a specific task. Furthermore, the inherent risk of AI must be addressed through smart and sensible AI regulation.
▶ 0:25:15Smart, well-crafted federal regulation will minimize the risk of AI, maximize its benefits, and prevent criminal and unethical use. These regulations are are needed to protect all Americans, especially small business owners. Small businesses are crucial to the technological progress of our country. Entrepreneurs use their unique perspectives, passion, and resourcefulness to support and grow our country's economy.
▶ 0:25:41That's why we must support our small businesses by educating them on this revolutionary technology, how to use it effectively, and how to avoid its risk. It is important that the federal government supports small businesses developing new AI technologies with safety and security in mind. It is vitally important that we do not leave small businesses behind as we race to enhance our country's competitiveness.
▶ 0:26:07I look forward to discussing how we can better assist our small businesses in this new era. With that, I thank the witnesses for being here today and I yield back. Gentleman yields back. We will now move to witness introductions. First witness is Mr. Jordan Krenshaw. Uh Mr. Crrenshaw is a senior vice president at the US Chamber of Commerce technology engagement center. Before joining the chamber, he was an attorney focused on environmental issues and consumer privacy laws. Mr.
▶ 0:26:35Krenshaw worked at Maguire Woods and also held roles in the Virginia Senate leadership, the office of the attorney general of Virginia, the US Department of Labor, and the National Right to Work Defense Foundation. He and both his under undergraduate and law degree from the College of William and Mary. I want to thank you for being here and look forward to our discussion. Our next witness is Anthony Caillum.
▶ 0:26:59Uh he is the president of M Apothecary, a locally owned small business serving customers across the Chicago area and through its online store, across the country. Uh the Caillou family has owned and operated the business for decades and under Mr. Caillou's leadership, it has grown from a neighborhood apothecary into a nationally recognized destination for natural health, personal care, and specialty products. Mr.
▶ 0:27:26Caillou Caillou earned his bachelor degree from Williams College, one of the great schools, I might add. And I look forward to your testimony and thank you for you being here today. Our next witness is Mr. France Hong. Mr. Hong is the founder and CEO of Boodle Box, a collaborative uh AI platform that gives small business educators and workforce programs secure, affordable access to multiple premium AI models in a single workspace.
▶ 0:27:55Prior to start starting Boodlebox, Mr. Hong served as an executive officer of US Army Special Forces Company during combat deployment in Afghanistan. Thank you for that. And I also he also served as associate counsel of the president under President George W. Bush. Uh and he has been uh on the founding teams of companies across the fields of law, aerospace, defense, government services, and AI. He earned a bachelor of science degree from the United States Military Academy at West Point.
▶ 0:28:25Beat Navy, right? And a law degree from Georgetown University. I want to thank you for being here today and look forward to our discussion. I now recognize Congressman Cisneros from from California to briefly introduce our last witness appearing before us today. Thank you, Mr. Chairman. It's my honor to introduce Mr. Chikagu, the head of innovation and strategy at the Capor Center, a tech policy group focused on making the tech industry more diverse, equitable, and inclusive.
▶ 0:28:55He is also the vice chair of the Maryland Higher Education Commission. Previously, Mr. Agu was the chief innovation officer at the last administration's labor department. While there, he led efforts to use big data and emerging tech to empower and protect American workers. Mr. Agu also has extensive experience in private industry, broadband access, nonprofits, and the education system. He holds a bachelor's degree from TUS, master's degree in education and public administration from Harvard, and an MBA degree from University of Penn Wharton.
▶ 0:29:25A lot of education there. Thank you for being with us, Mr. Agub. Gentleman yields back and thank you again all of you for being here today. Now before recognizing the witnesses, I'd like to remind them that their oral testimony is restricted to five minutes in length. That's a big number around here. Five minutes. Okay. Uh and if you see the light turn red in uh front of you, it means your five minutes has concluded and you should wrap up your testimony. If you continue to go, you will hear Okay? And it'll get louder and louder.
▶ 0:29:54So So I now recognize Mr. and I will say one other thing, too. You're going to see members going in and out. It's not because anything you said or did. We have other hearings going on. So, there'll be people coming in and out all the time. Uh, I now recognize Mr. Crrenshaw for his fiveminute opening
▶ 0:30:10Chairman Williams, Representative Cisneros, and distinguished members of the committee. Thank you for the opportunity to testify today on how artificial intelligence is shaping the future of small business. My name is Jordan Krenshaw and I serve as senior vice president at the US Chamber of Commerce where I lead the chamber's technology policy work including artificial intelligence, data privacy, and emerging technologies. For small businesses, AI is not a distant idea. It is already helping mainstream businesses save time, serve customers, manage costs, and complete compete with larger firms.
▶ 0:30:40Tools that once required large budgets and dedicated technology teams are now available to entrepreneurs, family-owned businesses, and small employers across the country. A retailer can use AI to understand demand and personalized marketing. A restaurant can improve scheduling and inventory. A professional services firm can automate routine drafting uh and research. A manufacturer can reduce waste and improve operations. In each case, AI is a force multiplier that helps small businesses do more with limited time, staff, and resources.
▶ 0:31:08The AI adoption trend is clear. According to the US Chamers's Empowering Small Business report, in 2025, 58% of US small businesses reported using generative artificial intelligence. That's up from 40% in 2024 and 23% in 2023. That growth reflects a simple reality. Small businesses adopt technology when it delivers practical value. One example is Henry's House of Coffee in San Francisco, a family-owned roster that has served its community for generations.
▶ 0:31:36Its owner, Hog Kbian, uses AI to improve online product descriptions, better understand customers, and solve production challenges. He even used AI to help design a custom system to address static uh in his coffee bagging process, which actually helped him save time on getting coffee outside of his uh e-commerce uh platform into customers. That is what AI adoption looks like on Main Street. Practical, entrepreneurial, and focused on solving real problems. AI is also creating workforce opportunities.
▶ 0:32:06The Chambers's data shows that 82% of small businesses using AI increased their workforce over the last year. These businesses are not using AI to replace people. They're using it to help employees focus on high high value work that requires judgment, creativity, and Small businesses are also investing in their workers. They need accessible training, strong partnerships, and policies that make adoption easier, not harder.
▶ 0:32:30It's one reason that the US Chamber and its foundation launched the Small Business Basics program recently to train 40,000 small businesses across the country on AI. We also support the AI for Main Street Act and the AI wise act, both of which would help put practical AI guidance and education within reach of more small businesses. At the same time, barriers remain. Small businesses face cost, technical data readiness, trust, and legal uncertainty challenges.
▶ 0:32:55The greatest policy threat in front of small businesses is a patchwork of state AI and privacy laws. A small business selling online should not have to navigate dozens of conflicting state rules simply to use common digital tools. Congress should establish a clear national framework for data, privacy, and AI. These frameworks should protect consumers, promote trust, preserve innovation, and provide strong federal preeemption. A first step is that Congress should pass the Secure Data Act.
▶ 0:33:23The Secure Data Act is a comprehensive privacy legislation that is based on a consensus of state privacy laws signed into law by 10 Democrat governors and 15 Republican governors that protects over 150 million Americans across the country. It would give Americans the right to delete their data and opt out of certain AI profiling. With regard to AI specifically, we should adopt a riskbased, technologically neutral, flexible, and proportionate framework.
▶ 0:33:49Congress should also support the foundations of AI adoption, workforce development, broadband access, reliable energy infrastructure, modernized permitting, government technology systems that reduce burdens on small businesses like uh streamlined permitting. The question before this committee is not whether AI is going to change small business. It already is. The question today is what policies can Congress put in place that encourage the adoption of this technology and further maximize this transformation to benefit small businesses.
▶ 0:34:16With the right approach that provides flexibility, certainty, and a clear standard, AI can help entrepreneurs grow, compete, and serve their communities. The chamber stands ready to work with this committee to ensure that AI works for Main Street, strengthens American competitiveness, and keeps the United States at the forefront of innovation. Thank you and I look forward to your questions.
▶ 0:34:36Gentleman yields back. I now recognize uh Mr. Caillou for his fiveminut opening
▶ 0:34:44Chairman Williams, Representative Cisneros, and members of the House Small Business Committee. Thank you for the opportunity to testify today about how artificial intelligence is helping small businesses like mine compete, grow, and create jobs. My name is Anthony Caillum and I'm the owner of MS Apothecary, a 151-year-old family-owned business in Chicago. We sell natural health, personal care, beauty, and fragrance products through our historic retail store, our website, and online marketplaces like Amazon.
▶ 0:35:14Altogether, we employ more than 60 people. When I started, I believe technology could help our unique neighborhood store reach people across the country and around the world. We launched our first e e-commerce website in 1997 and became one of Amazon's earliest small business sellers in 2004. Looking back, those decisions changed the trajectory of our business. Today, I believe artificial intelligence represents the next major leap forward for small businesses.
▶ 0:35:42AI gives businesses like ours affordable access to capabilities that were once available only to large corporations with multi-million dollar budgets. Many of these tools offer free versions or integrated into existing platforms we use. For example, Shopify's native AI can produce ondemand sales reports in just minutes that used to take hours to build. Over the years, we've accumulate accumulated decades of emails, spreadsheets, sales records, and operating documents in Google Workspace.
▶ 0:36:12We can now simply ask Gemini where something is stored and what it says, turning our historical institutional knowledge into something we can find and use instantly. In addition, AI helps us streamline repetitive administrative work so our employees can spend more time serving customers. Right now, we're building an AI agent to handle routine accounts payable work that takes one employee more than 20 hours a week. Hours we'd rather have her helping with strategic problems to grow our business.
▶ 0:36:41Like many small businesses, we don't have the budget to hire a full-time chief financial officer. But today, I can securely upload our financial statements into a AI tools like Claude or Gemini and receive a thorough analysis of cash flow, inventory, investments, operating expenses, payroll trends, and potential financial risks. Lastly, and perhaps my favorite example of how AI is helping us, we recently hired a national sales manager for Coobo, our small home fragrance manufacturing business. This is a new position.
▶ 0:37:11I knew what I wanted her to learn during the first few months, but I didn't have the time to organize everything into a structured onboarding program. Using AI voice transcription and writing tools, I talked through my ideas, and the AI turned them into a comprehensive knowledge guide along with a detailed 30, 60, 90day onboarding plan. She later told me it was the clearest roadmap she'd ever received when starting a new job. All these examples illustrate the larger impact AI has had on our business.
▶ 0:37:39It has helped make us it has helped us make better use of our employees time, strengthen our operations and grow. Since we started using AI, our headcount has increased by about 20% from just over 50 employees to more than 60 and we are still hiring. As Congress considers AI policy, I urge you to focus on measures that help small business adopt and benefit from AI powered tools. First, we need a clear, balanced national framework rather than a burdensome patchwork of conflicting state AI and privacy laws.
▶ 0:38:10Navigating different rules across all 50 states is difficult at best and expensive or impossible at worst. [snorts] Second, AI powered tools are just becoming affordable and accessible for businesses like mine. From features built into platforms like Shopify to standalone tools like Cloud and Codeex. If new rules make these tools more expensive or harder to use, it could push AI powered technology out of reach, once again making these capabilities available only to the largest companies.
▶ 0:38:38Lastly, Congress should invest in AI literacy. SBA programs, tax incentives, or workforce training initiatives could help small business owners and their employees learn how to use these tools effectively. Every employee who becomes proficient with AI becomes more valuable, more productive, and better equipped for the jobs of the future. Even in our business which is embracing AI, many of our employees lack formal training on how to maximize its potential. With the right support, more small businesses could help their workers use AI confidently.
▶ 0:39:08For 151 years, MRS Apothecary has adapted to changing technologies and markets. AI is the newest chapter in that story. It is helping us compete, innovate, hire, and better serve our customers. I hope Congress will pursue policies that keep AI tools accessible for American small businesses and make it easier for millions more to adopt them. Thank you and I look forward to your questions. Gentleman yields back. I now recognize Mr. Hong for his five-minute opening
▶ 0:39:37Thank you, Mr. Chairman, Ranking Member Cneros, members of the committee. Thank you for inviting me here today. My name is France Hong. In 1975, when I was a small boy, American forces evacuated my family from Vietnam. That single act gave me a new life. It let me graduate from West Point, serve as an associate counsel to the president, and later deployed to Afghanistan as the executive officer of a special forces company. It also let me become an entrepreneur.
▶ 0:40:04Over the past 15 years, I helped start a law firm, Fluitt, a defense services company, Mag Aerospace. And today, I run a collaborative AI company, Boodlebox. Today, more than 130 small businesses run a boodle box and over 100,000 students and faculty at more than 10,000 colleges use it to teach and learn. Together, the companies I founded employed more than 1,200 Americans. So, when this committee talks about small business, I have lived that journey multiple times.
▶ 0:40:33This committee has also has already done real work to help small businesses through bills like the AI Main Street Act and the AI Wise Act. That work matters. But I've learned something on the ground. Getting access to AI is no longer the hardest part. Basic AI is cheap and it is everywhere. A small business owner in Detroit can reach more AI computing power today for almost nothing than a Fortune 500 company could buy 5 years ago.
▶ 0:41:03The trouble is what comes next. The premium AI that delivers real results still costs real money. And a new divide is opening up between Wall Street and Main Street. There's a second problem, and it is bigger. The most powerful AI is useless if you don't know how to use it. Well, most small business today are handed the keys to a Ferrari and sent down the equivalent of a muddy road. Too often, what comes out only looks finished, and when someone has to redo it, it's what now we call work slop.
▶ 0:41:34on Main Street today and even more so tomorrow. The businesses that do not learn how to collaborate with AI will not just fall behind, they will fail. When I sit with small business owners, I hear the same things over and over. They're exhausted. They don't have time to learn new tools. They don't have staff to run them and they don't have budget to hire help. And they fear that once their data goes into AI tools, it stops being theirs. They are right to be careful. It's one of the reasons we built Budabox to be private by design.
▶ 0:42:02Not only do we have best-in-class security, but we never allow our customers data to be used to train AI models. In fact, we are one of the few platforms approved by the House's own chief administrative officer for use by your staff. Um, a small business owner deserves the same asurances. I have seen what this looks like when it's done right. The University of Michigan runs the Detroit Neighborhood Entrepreneurs Project. It serves more than 400 small businesses a year. More than nine of nine and 10 of them are minority owned.
▶ 0:42:32Last year they put Boodlebox in the hands of all those entrepreneurs with the coaching to use it well. One of them is Teaia Gaines who runs a company called GT Coffee. Through Bootbox, she picks the right AI model for each task. One to price her products, another to build her marketing. Uh as a result, her average order has grown from $38 to $47. AI let her grow before she had the resources to hire. She didn't outsource her judgment to AI. She amplified it.
▶ 0:43:00I'd like to leave you with three requests. First, fund the roads, not just the cars. Put training and real support behind the small business development centers that the committee has already strengthened with the AI for Main Street Act. Second, protect small business data so no owner fears that using AI means giving their work away. Third, take care that the rules meant to do good do not wall out the small.
▶ 0:43:28Compliance that only the giants can afford will hand this technology to big companies and lock out those who need it the most. Throughout its history, America has pro provided opportunities that let small business owners thrive and prosper. In this era of rapid change due to AI, it's our responsibility to ensure that those opportunities not only remain but expand. To do so, we must not just hand small business owners a faster car. We must help them build better roads. Better roads worth driving on.
▶ 0:43:58Main Street, I guarantee you will take it from there. Thank you. I look forward to your questions.
▶ 0:44:05Thank you. And gentleman yields back. I now recognize Mr. Agu for his fiveminute opening remarks.
▶ 0:44:12Good morning, Chairman Williams, ranked members, and members of the committee. I want to thank you for this invitation to testify today. And I will say as a former congressional intern who walked these halls two decades ago, it is a remarkable full full circle moment and uh very much proof of what's possible today uh in America. Um my my name is Chagua. I serve as head of innovation and strategy at the Kapor Center and through the Capore Foundation and Kapor Capital.
▶ 0:44:37The Kapor Center has backed more than 200 startups and fledgling businesses since 2011 that make healthcare, child care, education and housing more affordable today in America. Few topics matter more than the intersection of artificial intelligence and the small businesses that power this nation. If you remember nothing else I say today, remember this. The rise of AI carries both promise and peril. They are both real. Our shared responsibility is to capture that promise and mitigate that peril. For small businesses, the promise of AI is very clear.
▶ 0:45:06Generative AI and AI agents can remove many of the very barriers that prevent entrepreneurs from starting and scaling businesses. for market research to back office capacity to navigating regulation. Last year, Matt Rosenberg uh in California left his marketing job to open a Thai restaurant with his wife, a fourth generation cook from Bangkok when a storefront lease proved unaffordable.
▶ 0:45:28AI surfaced a rare home kitchen permit they never would knew existed, cutting their startup cost by more than 90% and turned dense regulations into simple checklists. This February, Bangkok Rush Thai Kitchen opened in Alama, California, uh about 12 minutes from our from from our home office.
▶ 0:45:45And when Matt Rose asked what he would have done without AI in this situation, he simply said, "I would have just given up." Matt's story captures the promise of AI, making entrepreneurable um entrepreneurship possible for all, not just for some. But the parallel is just as real and it's three-fold and it's important to name. First is trust.
▶ 0:46:06right now AI adoption and there are ranging uh definitions uh and estimates of adoption varies across small business and it's nowhere near where I think we want it to be and I think what we like to say at the kore center is uh technology uh moves at the speed of innovation but adoption moves at the speed of trust and right now that trust is still being earned uh by this technology in the marketplace secondly concentration uh if AI and the knowledge of how to use it is only concentrated in the largest firms the gap with with with main street is going to only widen.
▶ 0:46:37Lastly, uh and third, fitness for purpose. If these systems are unreliable or poorly governed, then small businesses are left to navigate these risks alone. Unable, as a former White House n uh deputy national cyber director Camille Stewart Glouster has warned, they wouldn't be unable to protect their intellectual property, their consumers, or their competitive advantage. So, the promise or the peril, which will we live into? It is easy to feel like this will happen to us.
▶ 0:47:04But at the KPAR Center, we like to echo President Kennedy where he says things do not happen, they are made to happen. And so for the future that we want, how do we make it happen? And we would say that there are four parts to that answer. First, we have to ensure that small business leaders have both AI literacy and practical guidance on where AI creates the most value. The bipartisan AI wise act and the AI for Main Street Act are promising first steps and we hope they are only the Second, we must help firms build AI ready workforces.
▶ 0:47:33And AI ready workforce is one that is not just AI literate, but AI native, fully comfortable working alongside these tools, but also strengthening their unique human skills that AI for for now cannot replace. Leadership, communication, interpersonal connection and judgment. Third, we need a national framework for AI governance.
▶ 0:47:52We believe common sense guardrails like those in the AI Civil Rights Act, which we endorse, build trust by preventing unlawful AI discrimination, requiring transparency for critical uses, and providing real accountability. We also need consistent rules of the road on security, privacy, and intellectual property so that entrepreneurs, and this is what's so important, can spend their time serving customers and not navigating AI risk. Fourth, government must be a partner and a pioneer, not a barrier.
▶ 0:48:19Governments should show the way by partnering directly with AI native companies who to solve societal ills and take down the bureaucratic barriers that 6% of these companies say are barriers to their scale and to them serving communities. If those partnerships could happen and just make a 2% improvement on many of the nation's biggest challenges, we could unlock over a hundred billion dollars of GDP for this nation every At the Kore Center, we believe that America is stronger when there are more entrepreneurs in it, not fewer.
▶ 0:48:47We must listen to them, make sure these these solutions reach them regardless of who they are and where they live. Uh if these solutions that you are debating here are not felt in places like my home county of Prince George's or El Paso, Texas or Appalachia or or the Ozarks, um we will not get the next generation of small businesses. And so I want to thank the committee for having me. I want to start that we believe this is and say this is a time of great promise and peril and that we believe this future will not just happen, it'll be made to happen. Thank you so much.
▶ 0:49:17Gentleman yields back and we will now move to the member questions under the five-minute rule. Uh I'll recognize myself for five minutes. Uh Mr. Mr. Caillou Merch Apothecary has been a fixture on the on the Chicago street corner as you said for 150 years, but through smallflower.com you're now competing nationally imagine a catalog of more than 14,000 products with a staff of around 50. And as a national retailer, we have entire departments handling that kind of merchandising, product discovery, and online marketing.
▶ 0:49:47Uh can you walk us through briefly AI has let your team take on uh that used to require the resources of a much larger company and where has it closed the gap with larger competitors and where does it still fall
▶ 0:50:04Yes, thank you for the question,
▶ 0:50:08you know, when you're running a small business, as I've done for about 28 years, uh, and I've worked in the corporate world, in the corporate world, you have departments who, um, can, you know, a department for every task. And in our business and in most small one, two, a small handful of people are wearing all of the hats across that would fall across a a much larger corporation.
▶ 0:50:31So for us, the way we use AI is often I I think of it as we're looking for ways that we can take the simple work uh and the tasks away that we have to do so that we can focus on the deeper work that helps our business grow. So that might be like the example I gave um matching bills to invoices which is obviously important and we can't exist if we don't do that but it isn't necessarily the best use of a human's time uh because we have a lot of other things to do. So that would be a simple example.
▶ 0:51:02And then in marketing, you know, nowadays we're able to test our marketing across the country to all types of segments. And then we used to sit there and think, should we test this ad for this segment in this part of the country or for this group of people? And it isn't really realistic to to do all the testing you need to do to succeed as a small business.
▶ 0:51:24So now we're able to use the AI to test across a lot of different creative and uh target customers and it's been a huge help for us. Those are two examples.
▶ 0:51:35Thank you, Mr. Krenshaw. Research shows that nearly threequarters of small businesses are now using AI, but the headline number surely hides a lot of variation. Uh question is how representative is that data across different sectors and regions and what separates the small businesses that are successfully adopting AI from the ones falling behind? Um as we look at the data um you know 58% of small businesses according to our survey nationwide are using generative artificial intelligence.
▶ 0:52:02Uh now if you look at the higher adopters that tends to be smaller tech firms that's around 70% adoption rates at that point. Entertainment companies that are smaller it's around 65%. But then when you dig down in the numbers, you know, you still have about half of retailers, restaurants, traditional brick-andmortar uh companies are using generative artificial intelligence uh to grow and thrive.
▶ 0:52:24Um in fact, had a chance to talk with the owner of a skincare company in the Coachella Valley uh who has been able to use generative AI to help her cut cost by sourcing products from places with lower uh tariffs or lower costs uh due to inflation. uh so she doesn't have to pass those costs on to consumers. Uh and so it really is a mix of the type of companies that are using this uh across all sectors.
▶ 0:52:49Uh but at the same time uh what we have found though that the the companies that are not using generative artificial intelligence uh about a third are worried about uh the regulatory uncertainty that's out there. Uh about one in five uh believe that they can't find the workers who are properly trained at this point to actually use generative AI and make it work for them. uh and then another third are are really uh you know still not sure how it will work for their business.
▶ 0:53:14Uh so it it's it's one of the reasons why it's so critically important uh that we uh establish uh clear national frameworks on the regulatory side like the secure data act on on data privacy uh create a national AI framework uh that's workable uh and we also have training programs in place that uh like the AI for main street act and AI wise act uh that that help uh small businesses really get their head around how to use small business respon or AI responsibly.
▶ 0:53:38Uh Mr. Hong, in the short time that I've got left from your work with Boodlebox, what does AI literacy really realistically look like for a typical Main Street owner and and their employees? And how can platforms like yours and the broader training ecosystem help them build those skills in a in a way that they supports running the business rather than becoming another burden on their time?
▶ 0:53:59Yeah, very briefly. Thank you, Mr. Chairman. I think AI literacy has three components. First, domain expertise. uh you have to know what questions to ask, what context to provide eval be able to evaluate AI outputs. Second, AI enablement knowing when and how to use AI. Not every problem requires an AI solution, but if it does, what are the tools and how do I use them responsibly? And third, I would argue that in this era of AI, the things that make us uniquely human are more important than ever. Critical thinking, collaboration, communication, creativity, and so those combination, those three things are what we need for today's workforce.
▶ 0:54:30Okay, thank you. My time is up. I now recognize the ranking member for five minutes of questions.
▶ 0:54:35Thank you, Mr. Chairman. You know, many of my colleagues and I were deeply concerned when we saw the president's executive order on AI published in December. There are several concerning elements to this executive order, but the most concerning is a requirement by establishing a uniform federal policy framework for AI that preempts state laws on AI as well as data centers. I support strong federal AI regulation and I think it's important that we have that.
▶ 0:55:01But I am worried that this EO punishes states like California, my home state, that are proactive and forwardinking enough to harness the benefits of AI while also creating guard rails against potential harms especially because this EO removes all state laws without putting any safeguards in place. Mr. AU can you share with a preemptive what can you share why a preempting state AI laws can be harmful without any strong regulatory safeguards?
▶ 0:55:31Sure. Uh Congressman, thank you so much for for the question and and I think for us the Kapor Center, we very much uh think the best way forward is a is a single national framework for AI governance. In lie of that, we do not think the best approach is no regulation. Uh we believe that right now if I cannot have national regulation then someone needs to step in and right now states are stepping forward to to create those safeguards.
▶ 0:55:56One can argue how optimal that is but we would say it is more optimal than having no regulation or protection whatsoever. Particularly for this topic that we're talking about today if you talk to many small businesses about why they may not be adopting is because they are unclear about what the rules of the road are and what protections exist for them. whether it be around their data, around their security, around civil rights, and having those rules, even if they're state-based rules, can actually be a tool to increase trust, which can increase adoption, which I think we've all said on this panel is to the benefit of small businesses in their
▶ 0:56:25Yeah. And like you said, the trust is the important thing, right? That's what's going to allow more people to use this tool. So, Mr. uh Agu, uh in your testimony, you said government should be a partner, not a barrier. Uh Rep. Jack and I introduced the SBA IT modernization reporting act to address challenges at the SBA in carrying out IT projects that support small businesses.
▶ 0:56:48I'm glad to hear it unanimously passed the House and now a Senate companion is being How how counterproductive can it be to small businesses if government agencies like the SBA improperly utilize AI?
▶ 0:57:03Again, Congressman, thank you so much for the question. And u it can be harmful in two ways. one uh the government is is in some ways an influencer and a leader for for so many other types of of industries across the country. Uh when there are uh misuses at the federal level, it it can potentially frankly incentivize and and almost uh create co-signing for those misuses in others. That's one. Two, when there are bad effects of those misuses, it creates a chilling effect for other folks who may want to partner with AI native AIdriven companies.
▶ 0:57:33And we think it's why the SBA, we want to be pioneering, but also strategic in terms of how they use it, so that they can actually show the way for what right looks like instead of highlighting what wrong can look like. And we believe if that happens, more small businesses will will be able to do more at every level of government and at every sector to improve their communities and improve problems around the country.
▶ 0:57:52All right, thank you for that, Mr. Hong. Um, I loved your example about the University of Michigan and how they're teaching companies how to uh use AI. Um, moving forward, as you said, literacy is is very important. What role should the SBA play in this, when it comes to literacy, AI, ensuring that small businesses know how to use these tools and use them properly to their benefit?
▶ 0:58:18Yeah, as we saw in places like uh Detroit, uh when you combine access to the tools with the training and coaching to use them well, that's when you see results. The SBA obviously has the resources and the reach to provide that type of training and resources. Uh obviously uh private industry has a role to play as well.
▶ 0:58:39But the reach of the government is much greater and particularly in you know urban America in downtowns in the in the far reaches of past the last stoplight of rural America. I think the SBA can really uh make an impact in terms of providing resources necessary for the responsible and collaborative use AI.
▶ 0:58:57Mr. Grinch, same question to you. What role should the SBA play in AI literacy?
▶ 0:59:03Uh, we believe that SBA should play a fundamental role in AI literacy. It's one of the reasons why we support the AI for Main Street Act uh and the AI wise
▶ 0:59:12All right. Thank you. With that, Mr. Chairman, I yield back.
▶ 0:59:15The gentleman yields back. I now recognize Miss Vanine from the Grey State of Texas for five minutes.
▶ 0:59:21Thank you very much, Mr. Chairman. You know, small businesses really do drive the economy. uh I'd say throughout the country, but especially in North Texas. Uh and AI seems to be giving um entrepreneurs new tools uh that were once out of reach for all but the largest of companies. It's helping owners save time, make better decisions so they can focus more on serving customers uh and growing their businesses.
▶ 0:59:42And on this committee, we are working uh we are working alongside the Trump administration to remove a lot of the barriers and to help small businesses to put these tools to work. uh Washington should be building on that progress uh without creating new uncertainty or unnecessary costs for our small businesses so that businesses of every size uh can benefit as American as America continues uh to lead in AI. Um Mr.
▶ 1:00:07Crunchshaw, AI is already helping small businesses uh compete with much larger firms. Washington needs to be careful as I mentioned not to regulate away that advantage. But at the same time u business owners need confidence that their customer data the financial information proprietary work uh are secure.
▶ 1:00:26So as we consider some of these privacy concerns what do you think that we should be looking at first and how can we address real risks without creating new compliance burdens that will make AI less accessible to mainstream companies.
▶ 1:00:42It's an excellent question and and you know there's a reason that Texas small businesses are using AI. 86% of them grew their workforce, the ones who used it. Uh, but they have concerns. 62% of Texas small businesses are worried about a patchwork of state AI and privacy regulations. Uh, uh, creating an environment where small businesses incur the costs and and don't reap the benefits and can't compete with larger companies. Um, I I think there really two things that that really can be helpful in creating a regulatory environment for uh for AI.
▶ 1:01:10One is passing national data privacy legislation. Uh the secure data act is currently uh in front of the ENC committee right now. Uh it is legislation that is based in large part on the DEX Texas data privacy and security act. Um and also uh along with 19 other states across the country as a a model for privacy legislation uh that would put um u consumers in charge of of their data and give them the right to delete it, correct it.
▶ 1:01:35Um but at the same time we also do need a national AI framework uh as well that that prevents a patchwork from emerging. Um, and I think as we look at that, one of the the first things we have to do is address where is AI already regulated. We need to do a gap analysis to see where uh the laws can already play a role in protecting people. Um, we already have laws in the books that protect against discrimination. Uh, privacy in over 20 states, uh, consumer protection laws in 50 states.
▶ 1:02:03We have an FTC act that protects consumers on a national level. Um and then at that point, uh there needs to be a a real discussion about a riskbased framework that's targeted and tailored for those areas of harms where the law has not caught up with the technology. Uh one example of that is um the take it down act which was uh signed into law which would have criminalized the distribution of deep fake revenge images and also given consumers a right to take that um those images off websites and off the internet.
▶ 1:02:30So that that's really where we feel a balanced approach is. Uh, and I think as we look at AI regulation, it's important to note um, you know, even if there is preeemption of specific AI regulations across the country, there's still a backs stop of these laws of general applicability like consumer protection, discrimination, and privacy.
▶ 1:02:47I appreciate the answer. By the way, I've asked a number of CEOs and founders of some of the largest AI companies, and that was the best answer I've gotten. Very complete and well thought out. I appreciate that. Uh, Mr. Dr. Hong, in your testimony, you discussed the importance of both proprietary and open models to small businesses, including the importance of open models due to cost. Can you kind of expand on the current state of the closed and open model ecosystem and how small businesses you use both types to operate and build their businesses?
▶ 1:03:14Yeah, to be clear, um I'm neither pro- open or or anti-open or proclosed or anti-open model. I believe in choice. Um choice provides competition. It provides the consumers the opportunity to use the models that are best. Um, we obviously are familiar with the closed models that are kind of on the frontier edge. There's also a number of open models that are very close. Um, they're driving on the cost to access these capabilities. Um, I built my platform on this idea of choice. You can choose what models you want to use and collaboration.
▶ 1:03:43So, you can get the best of models working together. Um, I think we need more competition in the sense of more access to both closed and open models so consumers can let competition does what it does best, which is produce the best products and services for
▶ 1:03:57I appreciate that very much. And with that, uh, Mr. Chairman, I yield back.
▶ 1:04:02Gentle lady yields back. And I recognize Mr. McGarvey from great state of Kentucky for five minutes.
▶ 1:04:07Thank you, Mr. Chairman. Thank you, Mr. Chairman, for having this hearing today. I think it's incredibly important for us to be talking about AI as it relates to small businesses. As you all have pointed out, AI is not something that's coming. AI is here. Uh we see it right now in Louisville, Kentucky. I see businesses who are successfully using AI in a lot of the ways you all described, right? Which is actually taking some of the more the tasks that take hours upon hours upon hours to do and then freeing up their employees to actually do the work that is central to the business.
▶ 1:04:35So see a lot of good applications have a lot of questions too, right? and see a lot of fear um from people who are afraid that AI is going to take their job entirely. Uh this is the kind of thing where like AI I mean could quite literally cure cancer um or or potentially be used to develop a a type of biological weapon.
▶ 1:04:55It's the type of thing where AI can be used to help spur creativity, but we don't want to see it ripping off the beautiful sound of someone's voice or taking the hard work and creative energy that goes into music and products and taking that and taking it away from us. How do we replace and and how do we how do we protect people whose livelihoods are easily replaced by AI? Um, how do we balance the needs for privacy with the needs for safety and security? And you guys talked about some of these things.
▶ 1:05:24So at the root of each of these questions I think is what I spend the most time thinking about when we think about AI right now. How do we make sure that AI is used for good and this is something that's I think is really important work of this body right now. And Mr. Crrenshaw, I'm going to start with you because you testified uh before the science committee all the way back in 2022 about the ethical development of AI. And I am not being facitious when I say all the way back in 2022.
▶ 1:05:51Um, when we talk about AI, 2022 was a very, very long time ago. Um, we could be talking about the beginning of this year and it might be a long time ago in in the way that AI is developing. Um, but I want to hear a little bit more about this and and I want to hear from you. What do you think the role of Congress is in ensuring that we develop AI that is ethical, that it that does good and is used for good? Uh, and tell me in any way if some of your views have evolved since 2022.
▶ 1:06:20It feels like a long time. Um, a lot has happened. A lot has changed uh since I testified in front of the House Science Committee. Uh, we have now seen uh the rise of new types of AI, for example, and more uh news about frontier models and agentic artificial intelligence and and the rise of chatbots uh over the last few years.
▶ 1:06:38Um but but I think one of the things that has been helpful uh in government um and was actually released during the Biden administration was the release of the NIST uh AI risk management framework which provides a great road map toward uh businesses of all sizes to uh examine and look at how they're using artificial intelligence to address risk to things like privacy and and other potential harms that could come out.
▶ 1:07:02Um it's one of the reasons why the risk management framework is iterative and meant to be uh re-examined uh to make sure that uh it is catching up uh with the technology and that's exactly what's happening right now. NIST is currently uh reviewing and and making changes to that risk management framework right now. Um it's also putting out profiles uh for different use cases for artificial intelligence. for example, uh uh just recently in the last year released uh a risk profile on generative artificial intelligence.
▶ 1:07:32It's now working on one on uh critical infrastructure. But it's also possible that that uh NISK could also uh examine ways for small businesses to more easily look at um uh how to use artificial intelligence in an ethical manner. So so that's a way government's already working. Um but you know, at the same time, you know, there are other things that we think that can be done. Uh, for example, Kentucky led the way on data privacy legislation.
▶ 1:07:56A super majority of Republicans and Governor Basher, a Democrat, uh, enacted Kucky's data privacy law, which has rights around AI. It has deletion rights. Uh, and and we think that that is the right model for the United States in terms of passing national privacy legislation, which would be a building block to getting us to understanding how data can be regulated. So then we can start looking at more of the broader AI pieces.
▶ 1:08:17No, thank you. I appreciate you saying that. This is one where I I wish I think all of us probably wish we had more than five minutes. And this is an issue I don't want to see get political, but you mentioned the importance of NIST. I have to point out that that this administration, the Trump administration has fired over 700 NIST employees. Uh and the fiscal year 2027 budget requests proposing 53% of that agency's funding. I think when we're talking about how do we use AI for good, how do we have this framework, an agency like that really is important for helping us develop those tools. Um I also appreciate you bringing up data privacy.
▶ 1:08:48Unfortunately, I am running out of time. Uh, but really appreciate you guys being here. Uh, appreciate the information you provided. This is something that is going to be so necessary for small business for everyone to be able to compete for us to develop the right framework. So, appreciate you guys being here. Mr. Chairman, I yield back.
▶ 1:09:03Gentleman yields back. I now recognize Mr. Alford from the great state of Missouri for five minutes.
▶ 1:09:10Well, thank you, Mr. Chairman. Good to see you and uh, everyone coming in today to testify on this very important matter. Artificial intelligence is quickly transforming the way Americans live and work. And nowhere is that more evident than on Main Street. And that's what we're concerned about here in this committee. What was once viewed as a technology only available to the largest of corporations is now becoming available and affordable for small businesses all over our great nation.
▶ 1:09:39Whether it's helping a manufacturer optimize production, allowing a family-owned retailer to better manage inventory, improving customer service, or reducing the administrative burden on small business owners. AI is giving entrepreneurs new opportunities to grow, to innovate, and to compete. At the same time, many small businesses still face significant challenges when it comes to adopting these tools. Business owners often tell us that they don't know where to begin. uh that they worry about protecting customer data.
▶ 1:10:10Many simply lack the time or technical expertise to evaluate the growing number of AI products entering the marketplace. We here in Congress, I think we have a very important role to play to make sure that innovation is encouraged, not hindered by unnecessary regulation, while also making sure small businesses have access to the education and resources they need to use this technology wisely to grow their businesses. Mr.
▶ 1:10:37Crrenshaw, uh, I was proud to introduce the bipartisan AI for Main Street Act along with my colleague, Representative Schultton. And this legislation would utilize the Small Business Development Center network to provide practical AI training, cyber security best practices, data protection guidance, and technical assistance. The House passed this bill by an overwhelming bipartisan vote because we recognize here in Congress that AI adoption is about more than simply providing access to technology.
▶ 1:11:07It's really about giving entrepreneurs the confidence to use it effectively. From your perspective, how valuable would this type of locally delivered techn technical assistance be? And what additional resources should Congress consider to make sure main street businesses can successfully adopt AI? Sir, uh well, first and foremost, want to thank you for your leadership and sponsorship of the AI for Main Street Act.
▶ 1:11:31Uh it's a critical uh step to ensuring that small businesses uh know the tools available to them, know how to use them in a proper and ethical manner. And it's a reason why it's so important as you look at at your home state, 87 87% of Missouri small businesses using AI have grown their workforce. So it's this is why we need to get these kind of tools in front of them. Uh so uh we think it's incredibly important that this bill uh passed in the Senate along with the AI wise act.
▶ 1:11:57Uh and then also SBA has the the funding it needs in order to successfully um uh promote AI adoption for small business.
▶ 1:12:04Thank you, Mr. CM. A big concern that I've heard from our constituents when I'm doing town hall meetings or coffees with people in small businesses. We're in those small businesses regularly when we're back in our district. Uh they're concerned about AI replacing employees uh especially entrylevel roles. In your experience, has AI replaced workers or is it more of a tool to allow workers to be more productive and more competitive for those small businesses competitive with the big guys?
▶ 1:12:34Uh, thank you for the question, Congressman. [clears throat] At MS, in my experience, it really is uh a competition uh advantage and it's I would say even more, it's a bandwidth advantage. And by that, I mean my in my almost three decades running a small business, I've never felt that there was a lack of opportunity. There was just a lack of time, money, and resources to tackle the opportunity.
▶ 1:13:01So for me it feels like what we've experienced is that using AI my team is able to go after bigger opportunities and help our business grow even the entry level
▶ 1:13:12Thank you Mr. Hone. Last question. What are the biggest obstacles the small businesses face after they decide to use AI? And what has helped your small business customers turn AI from a project into an everyday business tool?
▶ 1:13:26Yeah, I would say thank you for the question. Uh two things. First, uh, where to begin? Uh, AI, that chat window can seem like a lot of promise if you are an AI alert adopter. For someone who's not familiar with AI, that seems like an ocean you get lost in. So, uh, it turns out AI is actually very good about teaching you about AI if you use it right. It's one of the interesting kind of components of the tool. Uh, the second is we hear a lot about prompt engineering and the need to be able to ask AI the right questions. Um, in my opinion, what I've seen the most powerful use of AI, it's actually about context engineering.
▶ 1:13:57It's what you tell the AI before it gets started. And so educating business owners on how to use the context they're already sitting on, the customer data records, um your past performance, your past contracts, and how to feed that into AI in a way that's safe, secure, and private. That really unlocks the power uh and potential of
▶ 1:14:14Thank you, and I yield back.
▶ 1:14:16Gentleman yields back. I now recognize Miss McCyver from the great state of New Jersey for five minutes.
▶ 1:14:21Thank you, uh, Mr. Chairman, and thank you, ranking member, for convening this hearing today. I want to thank each of our witnesses for joining us today on a heat wave here in Washington DC. Uh so stay hydrated. Um artificial intelligence is often described as a tool to help small businesses become more productive and reach new customers, but its value depends entirely on the safeguards we put in place.
▶ 1:14:46Today's AI systems still reflect deep biases, raise serious cyber security risk, and operate with little transparency. While large corporations use AI to strengthen their advantage, many small businesses are left behind. We are not doing nearly enough to ensure that small businesses can access and benefit from these technologies on fair terms.
▶ 1:15:08AI's rapid growth depends on the massive expansion of data centers, which place heavy burdens on local communities, local electricity grids, water supplies, infrastructure, and people, especially in my district in New Jerseyy's 10th. These projects often move forward with little transparency or input from the communities most affected. That is why I introduced my AI data center site selection transparency act which would bring an end to surprise AI data center construction.
▶ 1:15:38If we are serious about encouraging innovation, we have to protect small businesses and communities that bear the cost of this Mr. Agu, you spent a lot of time talking about this in your opening testimony, which I can appreciate, about how AI has some potential to increase productivity, but it also has millions of America's workers wondering what will happen to their jobs due to increasing AI use. I want to elaborate a little bit more of what one of my colleagues talked about, Mr.
▶ 1:16:09McGarvey, about America, Americans in the workforce and how AI impacts that. Mr. Agu, how would small businesses or how should small businesses think about using AI to assist their employees rather than replace them?
▶ 1:16:25First of all, Congresswoman, thank you so much for the question and I think this is a place where I think history is helpful. If you think about in the past uh in the manufacturing space and you think about when someone running a factory was trying to figure out how to make the line run faster, the first place they went was they went to their employees and said, "You work on the line.
▶ 1:16:41Tell us how we can make this line 10% faster, 15% faster." and the work is generally delivered and I think the same ethos is is is is important here when uh business leaders are implementing AI one of the the people who must be intimately involved are the workers who particularly on the front line who are using it one because we that's ideally the types of workplaces that we want but secondly they have many ways the best ideas for how to make sure that that that that AI can be used to actually create the efficiencies that we want what can I take off my plate that'll make me better able to serve our customers grow our business serve our
▶ 1:17:12serve our community so I think that's number one include workers in that process and usually when we're talking about these fronts talking about 10 people 20 people uh many of which and in many ways these resources are already being created I think the second piece is and we've you've heard that I think across all the witnesses which is we have to be investing in their skills now again how do we make folks AI native so that they know how to use these tools the same way that we expect workers to be able to use Microsoft Office or Google Docs same way this is going to become that level of table stakes therefore how do we invest in that proactively now
▶ 1:17:42before it truly is a barrier to employment. I think the third piece, this is particularly true for small businesses and you heard this from my example of for Bangkok Rush Thai Kitchen, which is in many ways small businesses need to think about what was what is not happening in my business
▶ 1:17:57that AI could help make happen versus what is the thing that's happening now that potentially could could be done by someone else. And when you have again businesses of 10 or 15 or 20 people, there are lots of things that aren't happening because they just don't have the arms and the legs to to do it. So those are the three things that that that I would say include workers, make sure they're trained. uh and then thinking through strategically what's not happening in my business that AI could help make happen.
▶ 1:18:18Got it. Thank you for that. And many of you I think each of you actually talked about the need for AI literacy right for businesses. Uh in your opinion what investments in workforce training will be necessary to help workers adapt?
▶ 1:18:32Uh I think it's two things if it's in two buckets. One is just literally training folks on the tools. Uh again the same way that there are still classes that you can take on Microsoft Word and Excel. We we we and you have much content out there a lot of it either low cost or no cost which is good. I think we particularly need to make sure this gets to the frankly the most in need underserved workers and most underserved communities whether those be urban or rural because that's because those are usually folks who get that information last.
▶ 1:18:57I think the second piece and I say this as someone who used to work at the Department of Labor is how do we make sure that those human skills again leadership communication are emphasized and that's where things like work-based learning are really important. You learn those skills not by hearing about them learning about them but you learn them on the job. So apprenticeships hire to train uh hire train deploy um apparatus are really really critical. So more work-based learning and then more targeted uh knowledge of how to use these tools.
▶ 1:19:22Got it. Thank you so much for that. With that uh Mr. Chairman I'm out of time. I yield back. gentle lady yields back. I now recognize Miss King Hines from the northern Mariana Islands, the great nor the great islands,
▶ 1:19:34the amazing
▶ 1:19:35for five minutes.
▶ 1:19:36Northern Marianas.
▶ 1:19:38Um, thank you, Mr. Chairman. Mr. One of the things that resonated with me was your statement that um, excuse me,
▶ 1:19:46that the tool is only the car, the road is training, coaching, and support. And as the chairman has mentioned, I represent the Northern Marana Islands. And one of our biggest challenges often just our remote geography and our very limited access to technical expertise. Um so if Congress could do one thing to ensure that AI doesn't widen the gap for remote isolated communities like mine uh what what would that look like? What would that investment be?
▶ 1:20:17Uh thank you for the question. Uh I would say obviously pass the two acts that came up from this committee, right? The AI wise act and the AI mainstream act. uh that activates the you know SBA and the small business development centers to provide that training expertise and the government reach can reach all around the world right even to those uh territories and states u that are further a field so um I would say pass those acts um and and second continue to have the discussions continue to have these hearings and bring businesses small and large um the world is changing fast uh as
▶ 1:20:47some as one of the representatives noted if we have this discussion six months from now the world's going to be different
▶ 1:20:52yeah I agree um there's been a lot of uh I I think most of the panelists have touched on workforce development and you that's something I'm particularly interested in as you know this new technology continues to change on a daily basis if not by by minutes right and I just kind of want to flesh out some of your thoughts on on what we should be doing in terms of investing to be able to prepare our workforce for this new technology and I'll start with you Mr.
▶ 1:21:22Krenshaw.
▶ 1:21:24Well, I I think it's important to note too, in addition to AI training, um as we're looking at building out the infrastructure for uh AI, we're going to need training in that as well, too. Um you know, we've seen a rise in apprentichip programs for things like electrical work um to to help out build out the infrastructure that's necessary for for artificial intelligence. So, it's a reason why many of the companies we work with uh are are working to get uh people trained and those and those uh skills as well.
▶ 1:21:49Um, but it's also fundamentally important that we have small business owners trained on how to use this technology. And that's one of the reasons why uh the chamber launched its small business basics program. But also, you know, as we're looking at education, uh, we need to begin developing uh, programs where K through 12 has some degree of AI built into it, post-secary education as well too, uh, building AI uh, into those uh, curricula and to an interdisciplinary approach in a lot of ways and in higher ed.
▶ 1:22:17Uh and so it really is going to be an acrosstheboard um uh strategy necessary to ensure that we have an AI ready
▶ 1:22:25All right. Thank you. Mr. Kane, did you want to chime in?
▶ 1:22:28Yes. Um thank you. In my experience, you know, right now we are we're probably more advanced than most small businesses of our size in using AI. But that's because I happen to be into technology and myself and other leaders in our business have been pushing ourselves and our teams to do it. But even in our company, we feel like it's a challenge to get the rest of the company to adopt it.
▶ 1:22:54So I think that it is important to come up with uh programs ideally through the SBA or other means that help us train workers and let them experience um the advantages that come from learning how to use these tools. Well,
▶ 1:23:10Mr. Hong, did you want to say something as well?
▶ 1:23:12Yes, thank you. Um we are privileged to work with almost 170 colleges and universities uh on their teaching and learning with and about AI. uh what artificial intelligence needs is more actual intelligence and it's that combination that allows us to unlock its potential. Um if you think about the nature of work, right? Defining a problem, designing a solution, creating that solution, validating it. AI makes those middle two steps easier and cheaper than ever. But the key isn't those two steps. The key is what comes before. Am I using AI for the right problem?
▶ 1:23:42And what comes after? Does this solution actually work in the real world? And so workforce readiness maybe ironically enough isn't about learning how to use AI and the tools. It's learning how to apply them to the real problems and to solve real problems in the world.
▶ 1:23:57Mr. Agu, you have the rest of my time.
▶ 1:24:00I think the only thing that I would add here is more real world and you kind of heard it in the last answer which is again uh for those of us who are athletes at any point in our lives, you didn't read a book on how to do a jump shot or throw a football. you did it in the field with a coach and I think right now um AI is ironically showing the need for more work true in the field in the place of work training and learning so folks can learn to use these tools in real time classes are good real life experience is better
▶ 1:24:30in terms of learning how to use these tools because they're evolving so fast and so I think that is the only thing I would add here which is how do we make sure that people really young people are getting to use these in the field not just in a classroom
▶ 1:24:39thank you I yield back Mr. gentle lady yields back. I now recognize Mr. Tran from the great state of California for 5 minutes.
▶ 1:24:47Thank you, Mr. Chairman and ranking member, and thank you for our witnesses for being here. AI is no longer a technology of the future. It is already changing how Main Street does business. From helping a a uh family-owned business manage inventory to allowing a new startup to reach new customers, AI AI has the potential to make small businesses more productive and more competitive if we use it well. As someone who represents Southern California, I see both the promise and risks every day.
▶ 1:25:16LA County and Orange County are home to a booming AI and tech innovation ecosystem and also home to thousands of immigrant-owned small businesses that are worried about what AI means for their livelihoods. There's a broad bipartisan agreement that America should lead the world in AI innovation, but leadership also requires rules of the road. That's where Congress has fallen short. We still don't have a clear national framework for AI govern governance and that uh addresses safety, innovation, and equal opportunity.
▶ 1:25:46Indeed, access to AI is not equal. In many minority communities, access to high-speed internet, modern devices, digital skills, and language accessible tools remain uneven. This digital divide simply does not disappear during time of rapid AI adoption. It deepens. This is why AI literacy must come a na become a national priority.
▶ 1:26:07Small business owners should know not only how to use AI but also how to identify its risk, protect sensitive business data, recognize bias, and make informed decision about when AI should and should not be used. That education has to reach every community, including immigrant entrepreneurs who are often among the most innovative small business owners in the country. The largest corporations already have the resources to adopt AI, manage compliance, and pay for future subscription costs. Many small businesses do not.
▶ 1:26:38If we fail to close that gap, AI could accelerate market concentration and leave Main Street struggling to compete. We need to make sure that this doesn't happen. Congress needs to invest in AI literacy and establish thoughtful federal guard rails to ensure that AI becomes a tool for innovation and opportunity and not another barrier to success. Mr.
▶ 1:26:59Wang, given your background, we both know the opportunities that this country provided our families and we both chose to give back through military service before pursuing entrepreneurship. Your story is a powerful reminder of what the American dream can look like. In your testimony, you make an important important point that the challenge today isn't only access to AI, it's also readiness. You say many small business owners have been handed the keys to a Ferrari and sent down a muddy road.
▶ 1:27:28I never got that for Ari by the way, but uh can you explain how important AI literacy is for consumer uh and entre entrepreneurs? Thank you Mr. Tran and I appreciate your service as well given our shared heritage. Um let me just draw that analogy a little wider. So we tend to think of AI like a faster car. Let's just do whatever we're doing today and just do it faster and better with AI tools. But if you think about it, a car is not just the faster horse. All right? A car requires first infrastructure.
▶ 1:27:58It requires training. It requires rules of the road. A car requires um you to build the roads that can handle how fast the car can go. And most importantly, a car can go to places horses can't. And so in many institutions, organizations, public and private, we treat AI like a faster car. We simply do what we were doing before, just do it faster. I would argue the key to AI adoption and the key to AI in small businesses is to rethink the roads. to ask ourselves where do we want to go?
▶ 1:28:27And it may be a place we could not go to before. It may mean that a business that didn't have access to marketing tools because they couldn't afford it now all of a sudden does. It may mean that a business that couldn't expand um its market could. AI enables these new opportunities, but only if we build the better roads, only if we build the privacy, the data protections, only if we provide the training, and only if we ensure that the regulatory environment doesn't overly burden small businesses compared to large businesses.
▶ 1:28:54Let me follow up and ask you is AI literacy a challenge right now in this
▶ 1:28:59Yes, it is absolutely a challenge.
▶ 1:29:00Can I ask the re the all the panelists going down starting with Mr. Krenshaw if you agree?
▶ 1:29:11Absolutely. Yes.
▶ 1:29:12Okay. Then let me follow up with this. Should there be a dedicated funding uh federal funding to help bolster AI literacy nationwide? And let's go down the panel. Mr. Krenshaw. We should have public private partnerships that work together. Accept
▶ 1:29:28I'm not an expert on how it gets done, but I know we would appreciate it in our
▶ 1:29:34Same answer here. I I believe public private partnerships and from library where we can get AI literacy.
▶ 1:29:39Mr. Goo.
▶ 1:29:41Uh yes, but we will need public
▶ 1:29:43Okay. Thank you very much. And chairman, I yield back.
▶ 1:29:46Gentleman yields back. I now recognize Mr. Fuller from the great state of Georgia for five minutes. Thank you, Mr. Chairman. Mr. Wayne, and Mr. Krenshaw, you may weigh in on this, too. So, my my wife is a small small business owner. Um, I'm a I'm a prosecutor by trade. So, uh, if you remember the old caveman lawyer, your AI world frightens and confuses me.
▶ 1:30:10So, one of my concerns for small businesses on Main Street and the Appalachin communities that I represent is no different than the software you may use in in a business. you know, someone will sign a contract and then maybe a year, two years into it, maybe it's it's not delivering. I think you use the term um AI adoption, maybe it's not what they needed.
▶ 1:30:33um how can we create a uh an industry and framework to make sure that for that AI adoption on on Main Street that these businesses have the opportunity to become um get the AI AI that they need to actually grow their business and not get into a you know 24-month contract for something that's not responding to their needs. Uh thank you for the question. Um there's a lot of AI available. Uh some of it is free. Um a lot of it you can buy monthtomonth.
▶ 1:31:00Uh I don't think the uh the lockin is a problem. Um what is a problem is uh context right um there's going to be a fight for the context in which we provide our AIS and the context that emerges when we use AI and uh I believe that users should have the ability to move that context from plat platform. We we do that inside uh my platform in the sense of like you can share context between different models. So if today Claude is the better model you use that. if tomorrow Gemini is the better model, you use that.
▶ 1:31:30And so I think um being model agnostic in what you create is the right path forward for industry providing these types of tools.
▶ 1:31:41I think it's important too as as companies are looking at uh which AI to deploy. Uh they shouldn't have AI be the solution for its own sake. uh they should actually go in with an identified set of challenges they're trying to solve for and then look for uh the AI solutions that's best talented or best positioned and targeted for for solving those those issues. So uh you know I I think that's one of the reasons why it's so important that we do have facilitated AI training and education so small businesses can get smarter about what tools might be better for them.
▶ 1:32:08And I wanted to ask you about that Mr. Krenshaw. So, um, when you're talking about a public private partnership, because if I'm thinking about a small business in in one of my main streets, I mean, they don't have the time to think strategically and have big thoughts on like what what are those questions to help me a with literacy and then the inputs that I need for the AI to improve my my business.
▶ 1:32:29Uh, because they're worried about, hey, I can't even get someone to show up for a shift and that's going to, you know, ruin my mother's day, which happened to to my wife where someone didn't show up to a shift. I don't have the the certificate to go in and do her her shift at her business. She's got to fill it on on on Mother's Day. So for those business owners that don't have the time to think big thoughts about what AI is appropriate for them, what does that public private partnership look like? So you can get that AI literacy for a small business to say this is what I need.
▶ 1:32:59Well, I think that's one of the the great things about the way that the AI for main street act and the AI wise act are are put together. They complement each other. uh the AI wise act provides a lot of online tools and online availability of information. Uh the uh AI for main street act would uh work with SPDC's basically to have more of those localized trainings and localized education efforts as well. It's also one of the reasons why you know we think it's important that the business community itself uh takes a role in in providing training.
▶ 1:33:26We uh at the chamber have worked with local and state chambers of commerce to do trainings for local small businesses as well uh on on AI through the small business basics program. Uh but also too I I think it's important for individual businesses to be forward thinking. I mean I one example is Har Kbian who is the coffee shop owner I mentioned in my uh opening testimony uh who's using AI for cutting his marketing down by about 70%.
▶ 1:33:49Um, but what he does is he actually has an hour innovation hour uh with his baristas and his staff at the coffee shop to just go through how can AI help them and and kind of going through new AI solutions to get them uh toolled and ready even if they don't stay at his coffee shop to go on somewhere else. So, uh it really is uh different uh aspects of of the ecosystem when it comes to AI that have a role to play.
▶ 1:34:13Thank you. Uh Mr. Chairman, I yield
▶ 1:34:17Gentleman yields back. I now recognize Miss Goodlander uh from great state of New Hampshire for five minutes.
▶ 1:34:22Thank you, Mr. Chairman. Thank you to our witnesses for being here for this really important hearing. You know, one of the through lines in this hearing has been a very powerful point that each of you have made. Uh which is that getting access to AI is right now not the hard part for small businesses.
▶ 1:34:39The hard part is actually ensuring that a family-owned funeral home in Berlin or a home builder in Brooklyn or a daycare in Keen can actually put this technology to use and equally urgent whether five years from now or five months from now maybe is probably more accurate uh that small business will still have the leverage over the platforms that they built their businesses on. Um Mr. Dr.
▶ 1:35:08Wong, thank you for your testimony. Thank you for your service. You know, you've you your testimony has laid out for us a really clear vision of what doing it right looks like. Uh what closing the readiness gap looks like. And I I want to ask you, you know, our resources are limited.
▶ 1:35:25We're fighting tooth and nail for every dollar and and many dollars that this administration is withholding from from uh from my state of New Hampshire uh that could actually go to solve and close this read readiness gap um with limited resources. How could we scale up a model like the one you describe in your testimony, the Detroit Neighborhood Entrepreneurs Project? How could we scale this up? What would be your first concrete steps that we could take to scale up this model?
▶ 1:35:55So I believe in public private partnerships. A lot of the work we do at Budabbox is with colleges, universities. They are the front door to democracy. Uh today's college students are tomorrow's workers. Um obviously state and federal funding and governments, you know, are intertwined closely with higher education. Um you know, I would like to see greater support for AI literacy at the place where it it counts, I think, the most, right, during the formative years of someone's development just before they enter the workforce.
▶ 1:36:22Thank you. I appreciate that. I did want to ask about this. You make the powerful point that we got to fund the roads, not just the cars. To me, part of the road is connectivity itself. I represent one of the most rural districts in the country and access to high-speed broadband is for many of my small businesses a core barrier to the adoption of AI tools. Can you speak to this in and in the rural small businesses you've worked with?
▶ 1:36:50Um, and can you can you just you agree that connectivity is part of what it is going to require for us to build the roads?
▶ 1:36:59Yeah, absolutely. Representative, um, in my written testimony, I talk about um, a rural family in Oklahoma. They live well past the last stoplight because they have good connectivity. They're able to access AI. They're able to have a marketing team, a development team, a business development team that is AI powered that they would not otherwise have access to. But obviously if they can't access the AI, they can't have access to those potential. And so connectivity is really a fundamental foundation for this access to the AI
▶ 1:37:28It is. And I just want to say the the broadband equity access and deployment act, the bead program has been so mission critical there. We've got $21 million that's currently being withheld from the state of New Hampshire that could go to exactly this problem. Um Mr. ago. I want to ask you, thank you for your testimony and for the powerful points that you've made about the role of competition and the the real perils that small businesses are up against.
▶ 1:37:56Um, before I came here to Congress, I was an antitrust enforcer. So, I think a lot about what um what the scales could have looked like and could quickly look like for small businesses. Can you talk to us about examples that you've seen in your work of AI applications that have actually made small businesses better able to compete with large dominant companies?
▶ 1:38:21Sure. First of all, thank you so much for the question and I think um let me step back for a second to say I think I've seen two things that have helped this. One um you hear a lot about transparency for a small business own to be able to know is this tool doing what I need for it to do. That is hard right now because the technology moves so quickly. So being able on the front end to know what you mean to be have solved and then actually able to track them on a regular basis is so important because that's the first step to know when I need to switch.
▶ 1:38:46I think secondly when you talk to get the answer to your question where we have seen the biggest advantages in a couple areas first uh qualified lead generation meaning if I'm a business who sells by getting on the phone and trying to sell u my wares getting a qualified list of who those folks might be versus just picking up the phone book super helpful using to tools like LinkedIn that's one secondly back office capacity whether it be accounting tax compliance things like that being able to track all that and make sure that I'm I'm proactive versus reactive I think the last piece marketing.
▶ 1:39:16So many small businesses don't have a website. So many small businesses are not advertising in any way, shape, or form because they can't hire a designer. AI can make those things happen.
▶ 1:39:25Thank you. I yield back.
▶ 1:39:27Gentle [clears throat] lady yields back. I now recognize Mr. Muer from the great state of Pennsylvania for 5 minutes.
▶ 1:39:33Thank you, Chairman. Uh thank you to our witnesses. Um certainly a lot a lot a lot to talk about and my apologies. We have other uh committee responsibilities uh many of them today. Uh Mr. Cure, I'd like to start with you, please. Um, you know, in business, I used to look at at technology in the early 2000s as it's either it's either a weapon that is used against you or a weapon uh that you use to benefit your company. AI certainly fits into that category and then some.
▶ 1:40:00Um, what what are the challenges small business facing as far as um you know, of course there's privacy, uh, there's, you know, cyber ransom challenges. I mean the the onslaught, the competitive nature of AI, whether you're there or not, uh and the usefulness of it. So maybe just give us a a quick uh 30 seconds on what what what do you main feedback you're getting from your uh
▶ 1:40:26Uncertainty in the regulatory environment is a major barrier. Um it's the reason why 57% of small businesses uh in uh Pennsylvania are worried about a patchwork of regulations that leads to
▶ 1:40:38Okay. Um, you know, I think examples of that are we have a u we have a good base in in about 20 states for data privacy legislation, but we see other states that are operating on the margins and trying to take more extreme approaches to regulation. Um, we've seen those states, for example, uh, craft legislation with the intent of only having their state attorneys general regulate to provide some certainty.
▶ 1:41:00Uh unfortunately, we've seen instances where the trial bar, for example, has gotten creative and used laws never designed for consumer privacy uh to have gotcha lawsuits uh for basic things like collecting cookies. So, if if you're an auto dealer who has a chatbot just trying to get in touch with uh you know, a potential customer, we've seen auto uh firms that have been sued uh with smaller ones for uh with demand letters from people who were never even customers uh using these invasion of privacy statutes that had nothing to do with consumer privacy.
▶ 1:41:29Uh we've seen hospital systems uh even sued under this uh as well in Pennsylvania. Uh and it's it's why we need reform uh of of this type of of litigation abuse that we're seeing across the country. Um and then it it really goes to the heart of why we do need a national AI framework as well. If we have a state-to-state patchwork there, it will be highly problematic because only the larger companies are going to know how to actually comply and compete and it's going to leave small businesses in the dust.
▶ 1:41:54So basically one of the most useful tools we've had in a long while to create productivity and efficiencies and effectiveness. One of your biggest challenges is government involvement and intervention on the primarily at the state level.
▶ 1:42:08I think at this point if you look let's look at data privacy for example uh there was a great survey that came out from ITIF that showed that if we have a 50-st state patchwork of data privacy laws uh it will cost the American economy $1 trillion over 10 years 200 billion of that is going to be incurred by small businesses.
▶ 1:42:25Uh another example I think if you look at the California uh consumer privacy act uh there was a rulemaking that came from that uh that was finalized just last year uh regarding automated decision systems regarding cyber audits regarding privacy risk assessments. Uh the uh agency took a oneline part of the California Privacy Act and turn it into a half billion dollar AI rulemaking. When you combine that with the full rule making with things like privacy, uh the agency estimated it was going to cost small businesses $16,000 annually to comply.
▶ 1:42:56Wall taken serious of what you're saying here. All right. Thank you. You mentioned about Basel 3 basically that access to capital is will be a very very important right now particularly with what we passed in the uh working family tax cut portion of the BBB uh where we um have better u bonus depreciation and better better tax incentives for investment. So clearly that's something that's important. I'm I'm going to I'm going to move on to Mr.
▶ 1:43:24Com um in your business. Congratulations Mertz uh uh how's AI uh helping you and and I was interested the way you said how it now allows you to compete more so against the so-called bigger guys. Maybe you can talk about that a little bit.
▶ 1:43:41Sure. Thank you, Congressman. You know, uh I always there's a mantra that when you study small business or you look at the business leaders out there, they always say to work on your business, not in your business. And for most small business owners, that's just like
▶ 1:43:56it's a dream, but it's never reality. And I feel like in the past year since we've been using AI more heavily in our business, it's the first time where I feel like our leadership team has been able to work on the business. And when you're a small business owner, so much of your thoughts, what I what the reason I lose sleep is how do we make sure something doesn't happen because it's small and you feel like, you know, a pandemic can come along uh anything can come along and take that away.
▶ 1:44:24the AI investment making cost prohibitive or or you you can
▶ 1:44:27not at the moment. The privacy uh laws are the biggest concern about using it right now.
▶ 1:44:34All right. Great. Thank you, gentlemen, very very much. Uh chairman, I yield
▶ 1:44:38Chairman yields back. I now recognize Dr. Morrison from the great state of Minnesota for 5 minutes.
▶ 1:44:43Thank you very much, Chairman Williams and Ranking Member Velasquez for holding this hearing and thanks so much to the witnesses for being here today. Over the past years, we've been talking about this all morning. Uh we've seen rapid developments in the field of AI. Um AI tools of all kinds are readily available and some have the potential to support entrepreneurs. But I do worry that as tech giants race ahead to push out their next AI model, there's insufficient attention on the actual utility of this technology to our small business owners and frankly to the American people.
▶ 1:45:13Too often AI development narrowly focuses on finding out what AI is capable of rather than answering the question, how can this technology be used to improve people's lives. Now we find ourselves in a situation where many people are asking who even wanted this and will this technology help me or will it actually hurt me or maybe even kill me. Um Mr. in your testimony you highlight that technology moves at the speed of innovation but adoption moves at the speed of trust and that trust has not yet been earned.
▶ 1:45:42I think this is a really important part of this whole conversation. Um, you go on to highlight that this lack of trust is reflected by the America majority of Americans feeling that AI's risks outweigh its benefits and the slow adoption of generative AI by American small businesses. Um, you know, Rep. Tran touched on AI literacy. Um, Mr. Goo, can you share steps that tech companies could take to build America's trust in First of all, thank you so much for the question.
▶ 1:46:13I think is this is a really important point. I think I I go back to a point that's been made, I think, repeatedly, which is one thing that creates trust is beyond any company, but it is having a national set of frameworks around what are our rules of the road around AI from security to privacy to bias to civil rights having a single set of rules that we all know, understand, agree upon and and can trust it.
▶ 1:46:35I think secondly um having more of collaborations between uh the public and the private sector like these companies to solve really large issues again around things like uh some examples here are things like the claude core basically claude uh anthropic working with with nonprofits to to teach them how to use these tools for the benefit of the work that they do in their communities. Small example but a good one and we'd love to see more there in that vein.
▶ 1:47:01I think lastly um helping small businesses pick the tools that are right for them even if those tools are not their tools. I think that is really important so that so that consumers and small business leaders have those choices is so important. I think right now you have to make sure it is tempting to believe that everything is a nail and that is amendable to the AI hammer or particular AI hammer and helping these small business leaders pick the exact hammer for the exact right nail is so important. It might be claude. It might be chatgpbt.
▶ 1:47:32It might be a midjourney. It might be a notion. But how do we basically figure out what those tools are and putting that agency in the hands of small business leaders I think is really key. Last point, um, small business leaders also reflect the feelings of their workers. So the more that workers become more comfortable again are involved in these decisions, figuring out how it can benefit the business, the better off their employers are going to be and more trusting their employers will be with these companies.
▶ 1:47:55Thank you so much. And I am sympathetic to this idea that we have a if we have a patchwork of state regulations across the country that creates all kinds of Could any of you give two specific examples of what a federal framework would look like for regulating AI?
▶ 1:48:15One of the notes I think important to note 68% of small businesses in Minnesota are worried about a patchwork. And one of the solutions can come out of Minnesota is copying Minnesota's data privacy law and making it the national standard. Uh and so that's a
▶ 1:48:29love that answer. [laughter]
▶ 1:48:31Sorry to interrupt.
▶ 1:48:35I'd say one piece that is not talked about enough is intellectual property. Meaning uh when lot of intellectual property was created back when it was this use was never thought of, never considered. So whether I'm Universal Music Group who regulates, you know, 30 has 30% of the world's music or candidly I am a bluegrass artist from uh from K from Kentucky.
▶ 1:48:57I may never have thought of what happens if my music if my music is you used by um this and we need a framework to govern how I say it can be licensed, how I'm compensated, how I want to be compensated and by whom. right now we just don't have a framework for this because back then we didn't think that this would ever this would be possible. So that's one example of an issue that's under uh talked about that is really really key and a big trust
▶ 1:49:21Thank you. One I'm running out of time but one quick last question. You know before I ran for Congress I worked as an OBGYn um for many years. So as you might imagine patient confidentiality is critically important. Um and one of the pieces of AI regulation that I think is crucial is that data privacy question. Um, and I think it's something that people are worried about too. Um, and I know small business owners also have concerns about it. Um, how can we protect patient privacy? Mr. Agu, do you have any thoughts on it?
▶ 1:49:51I know I'm over time, so I'm looking at the chairman to
▶ 1:49:55Sure. Um, I think candidly, uh, we already have a lot of health privacy laws and I think the question is how do they intersect with with this new technology? And so I think that's the question. and how do we take existing regulations and expand it to this new technology and I'll stop there uh out of respect to the chairman.
▶ 1:50:11Thank you and thank you Mr. Chair. I yield back.
▶ 1:50:13Gentle gentleman yields back. I now recognize Mr. President for from P great state of Pennsylvania for five minutes.
▶ 1:50:20Thank you Mr. Chairman, ranking member and to all the witnesses for being here today. Over the past few years, artificial intelligence has become a part of nearly every corner of our economy, and small businesses have increasingly embraced these tools to help them compete and to grow.
▶ 1:50:38Whether it's summarizing research, analyzing data, streamlining customer service, or organizing documents, AI has allowed small business owners to compete in minutes, which used to take hours or even days. But there's no question these technologies can make businesses more efficient. But we continue to adopt AI. We also have to make sure that those gains don't come at the expense of American jobs. We need to make sure employees have the training and skills to use these tools efficiently and AI enhances their work rather than replacing it.
▶ 1:51:08So I'm going to my first question uh for Mr. Caum. Uh congratulations to you and your family and everyone at MS uh apothecary on celebrating 150 years of service to your community.
▶ 1:51:22My question would be at what point did you decide to take the initiative to start looking at different AI implementation and something that I hear from a lot of small businesses and coming from a small business background was sometimes you know the local plumber heating contractor doesn't even know where to begin.
▶ 1:51:42So at what point for you in your business did you decide that hey in order for us to scale in order for us to grow we need to be thinking outside the box and finding ways to implement some sort of AI technology.
▶ 1:51:56Uh thank you for the question congressman and for your Um you know I I came to the retail and uh small business this industry through the dot rise in the 90s. So I've always been into technology and so I may be a bit rare in that sense that I view it as my part of my role in the business has always been to stay ahead of the technology curve.
▶ 1:52:23So I've been following it for a few years now at least and I've been the person testing it out on the weekends and at nights to see what it could do and then bringing it to the team and saying look at this. So, I feel like for me, I I happen to be personally interested in it, but I know that that's not most of my small business friends, the other business owners on the block. So, I think um you know, we do need support in that sense.
▶ 1:52:50So, you've been around for 150 years. What generation are you in the business?
▶ 1:52:55I I'm the fifth owner in the second family to own the business.
▶ 1:52:58Wow. So,
▶ 1:53:00so what was the first piece of technology that you implemented that was in the ecosystem of AI? Was it just utilizing language models to respond to inquiries or to summarize research? What was like that that initial barrier that you overcame?
▶ 1:53:16Yeah, you know, we used some early language models around um coming up with variations like we would write marketing copy copy and then come up with 10 15 20 variations on that and that was many years ago but I would say chat GPT uh in the past few years was the first model that I started using um kind of you know as an experiment on my own.
▶ 1:53:39My last question would be and then before I divert over, have you actually needed to bring anyone on that had experience in managing AI models or has that been something that you've been able to sustain with with inside of your
▶ 1:53:55So far we have sustained it inside of our operations and you know the models are getting to the point where when we don't know something they can kind of often guide us on how to do it. So I call us sort of like you know pseudo techies in a way but um I think we're at the point now where we would like to bring someone in who has more experience because we see so much opportunity
▶ 1:54:16I think that's what we need to be curtailing some of our our educational opportunities and teaching the next generation of workforce how to work with inside of these parameters. Uh but also being very conscientious of we're not looking to replace jobs. We're looking to add jobs specifically to the skill set relating around artificial intelligence. And I think when you say AI, it's a very broad spectrum of innovation that could be implemented. So, I'm going to pivot over Mr. Hong.
▶ 1:54:44Uh, what are the biggest challenges you see small businesses facing when they begin implementing AI? And what kind of advice could I take back to some of my small businesses that are right on the cusp of considering to take the leap? uh just to kind of give them some asurances and relieve the aida about taking that initial that initial plunge.
▶ 1:55:06Yeah, thank you for the question. I'll try to be brief. Uh the the main two challenges are first just where to get started. What are the use cases? Uh AI can do so much. So what's the first thing? And I always advise businesses to start with the problem that is most at hand. What is the thing that bothers them day-to-day? That's where to get started. The second is AI is only as good as the information you put into it, the context. A lot of businesses are fearful of putting their proprietary data in. Um, that's unfortunate because that's how you get the best results from AI.
▶ 1:55:33And so the advice I would suggest you take back is get started and find a platform where you feel the privacy and security allow you to use your data. Thank you. Garbage in, garbage out. I
▶ 1:55:45Jim yields back. I now recogn recognize Miss Sculton from the great state of Michigan for five minutes.
▶ 1:55:51Thank you, Mr. Chair. Uh, and Madame Ranking member. Thank you to our witnesses today. Our small businesses juggle a wide range of responsibilities and tasks every day to keep their operations running. Uh while AI offers tools to help streamline many of these needs, we know that many entrepreneurs simply don't have the time or resources to utilize this technology to its fullest.
▶ 1:56:15That's why I was so proud to introduce my bill, the AI wise act, which directs the Small Business Administration to create learning modules for small businesses to help them navigate this emerging technology. I'm pleased that my bill passed the House earlier this year. Uh we're of course hoping to have it signed into law. Mr. Agu, how would uh an SBA resource to help entrepreneurs think about AI usage uh more effectively um how how would that benefit small
▶ 1:56:42Uh it would benefit hugely. one because again we talk about kind of the promise and the peril uh you've heard I think from [clears throat] a number of other witnesses you need certain tools and means to actually capture that promise and and the first thing is knowledge so knowing again where I can apply this most in my business should I apply it toward my inventory issue should I apply it toward my back office should I apply it toward my marketing first getting guidance on where to start with that I think is really helpful one I think two small businesses are unclear okay how do I introduce this with my workers uh that is a that that that is a hard part of this particularly when
▶ 1:57:13you're when you were in small business because there is fear, there is lack of knowledge, there is uncertainty. What does that change management look like? And then I think the third piece is how do I monitor these tools over time? Um the ubiquity of these tools, the amount of these tools is only useful to the extent that I know enough to switch when I need to.
▶ 1:57:30Therefore, what am I doing to monitor these tools over time to make sure, hey, you know what, actually it is chat GBT here versus Claude or I need to use a perplexity here versus a a another tool or there actually isn't a gap, right? there is a gap right now in terms in the market for what it for what I need. I think those are three specific areas where to apply, how to do change management, how to monitor over time where resource like this could be really really valuable.
▶ 1:57:50Yep. Yep. Um, West Michigan is home to our regional small business development center, SBDC, uh, which served over 500 clients last year and spurred $10.4 million in capital formation. These centers are foundational to the success of our local economies. And on a national level, America's SPDC's partnered with Google AIU to facilitate training SBDC staff across the country to assist clients with using AI in their operations. Not sure if you're you're familiar with that.
▶ 1:58:18Michigan now has the second highest number of staff trained in AI literacy in the country. We're super proud. Um Mr. AU again, how important is it that SBDC's receive full federal funding and support these kinds of activities? And I'll actually open it up to the whole panel. I saw some heads nodding. Uh, you know, feel feel free to weigh in uh on usage there.
▶ 1:58:38I cannot overstate how important the resourcing not just of the funds but also the people because in the end what we find is most small business people, they want to talk to people. Yeah.
▶ 1:58:46About how to better to use these tools and that does require resources.
▶ 1:58:50Additions, additional comments or or thoughts on that. I would echo that. I I think our team often times it's not enough to just keep hearing how good it is or watch this video. I think it is it's having some structured format and people who demystify it and provide a doorway in. And my experience is that once you give someone that doorway, then they can actually go really far.
▶ 1:59:15Yep. Um, in my remaining time, I I want to talk uh about um a different bill of mine, the SBI Adoption and Reporting Act. It's really important for small businesses to have the tools they need, right, to use AI effectively. Uh, but the federal government also has the opportunity to utilize AI for its own purposes and its own operation. The SBA is no exception. from evaluating contracting bids to improving research and programming.
▶ 1:59:42Uh the government can take steps to ensure AI is working for Main Street in every single way from within the administration itself. That's why I introduced the SBA AI adoption and reporting act which would direct the SBA to implement AI use disclosure policies and procedures as required by law and recommended by the Government Accountability Office. How could this help address small businesses concerns about the government's use of AI?
▶ 2:00:08Those people who might say, "I was weeded out because I didn't fit the algorithm, you know, uh or or they have concerns about that." Any any reflections from your work? I'll open it up to the panel
▶ 2:00:20Well, as the um beneficiary, our business is the beneficiary of an SBA loan. uh that almost didn't happen at one point because of some of the um challenges with the banks and everything. I would say that my experience would be having transparency and help along the way to make that an easier process would have been a huge help for our business and thankfully it did work out. But I'd like to see more businesses get that opportunity that we
▶ 2:00:48Heard. Heard. Thank you. Thank you. Well, thank you again, gentlemen, for coming uh for your time and most importantly for your incredible work uh in small businesses across the country. Thank you. Yield back.
▶ 2:01:01Gentle lady yields back. Right now recognize Mr. Jack in the great state of Georgia for five minutes.
▶ 2:01:06Thank you very much, Mr. Chairman. And I'd like to begin my line of question by recognizing someone in the audience with us. Alex Mueller is in the audience with us from Houston, Texas, the great state of Texas. So, I'm delighted that she had a chance to get to see how committee processes work. And I would like to start my first line of questioning with uh Mr. Cayam.
▶ 2:01:25And um you know, a question I have for you and it's something that I hear a lot within my district is there's a natural fear that AI could replace workers and um especially within entry- level roles and I'm just curious in your own experience, has AI replaced people or has it helped you bring more people in? Um what's your Thank you, Congressman. Um, in our experience, AI has helped us bring more people in. It's created more opportunity.
▶ 2:01:53So, I look, you know, AI right now is the this new um exciting potential big advantage for small business. But ultimately, it will be software and it will become a commodity and I think humans are the durable advantage that you have. So to us, we we look at it as way as a way to or as a method and a set of tools to help us increase our capacity and then bring more people on.
▶ 2:02:19And when we're serving our customers better and in a specialty business knowledge, we we speak I think 10 languages on our staff and have experts across various fields including pharmacists, makeup artists, uh certified herbalists, you name it. So that is not going to be replaced. People like to talk to people. So we see it as a creator of opportunity.
▶ 2:02:41Thank you very much. Mr. Hang, same question to you.
▶ 2:02:45Yes, if you're a small business of one to five employees, there's not many people to replace, right? AI actually enables you to grow your business faster and employ more people. Um I also like to note there's this interesting concept called the Javins paradox. Uh what it says is if you look a broad enough time horizon and you look across enough industries, technology ends up creating more jobs over time than it replaces it. technology will replace jobs. Um, but as we saw past technological revolutions, right, new jobs emerge.
▶ 2:03:15Um, I have an 118 and a six-year-old. I think I heard the statistic that 66% of the jobs that a 10-year-old's going to have don't even exist yet.
▶ 2:03:23Thank you very much. Um, Mr. Dr. Krenshaw, I'd like to speak specifically about an initiative um specifically the President Trump's SBA in collaboration with Perplexity AI recently launched, as you may be aware, the Main Street AI accelerator, which offers computing credits to small businesses that hold SBA back loans. Do you believe that a program like that is the right model uh when approaching the AI gap? And what would make uh assistance in this capacity most useful to owners who need it most?
▶ 2:03:54We think it's incredibly important that uh there's AI training uh and AI accelerators. So I I think um this is an opportunity where uh small businesses can grow. Um you know I think and I want to really kind of echo what the points of of some of the other panelists is that you know as we look at AI AI uh is doing much like what we saw at the turn of the 20th century with the turn of the 20th century technology brought about an industrial revolution.
▶ 2:04:18Well, now we're seeing technology create an entrepreneurship revolution where small businesses and small companies can get off the ground and run uh and and grow and and 82% of small businesses in Georgia grew their workforce, the ones who are using artificial intelligence. Uh this is confirmed in many ways too. We've seen some great studies as well by RAMP uh came out recently that showed that the more companies are investing in artificial intelligence, they're they're growing their headcount at faster rates.
▶ 2:04:44And then and then finally, I think it's really important to note too that the Federal Reserve uh has put out really great data over the last few years and and as you see the number of small business applications over the last few decades was pretty stagnant up until about 2020. That number skyrocketed after 2020 and many thought that that was in large part due to COVID and and and people kind of starting out on their own.
▶ 2:05:06But those numbers have continued to go up post uh introduction of uh chat GPT and generative artificial intelligence and that's been a sustained increase in new business starts and and and so when you look at all of that I think we're optimistic about the impact of AI but it's important that we also have training and accelerators for small businesses uh to ensure they reap the full benefits.
▶ 2:05:27Wonderful. In our closing 40 seconds, you mentioned Georgia. both my colleague [snorts] Klay Fuller and I represent uh districts that in some places need broadband access. Could you speak to how imperative it is that we're able to ensure that all of our constituents across the country have access to broadband internet that enables them to potentially enjoy some of these
▶ 2:05:46Uh broadband is essential to operating in the 21st century economy. Uh as a small business cannot compete without access to connectivity and to data.
▶ 2:05:54Wonderful. Mr. Chairman, I appreciate you convening this hearing and I yield back. Uh, gentleman yields back. I now recognize ranking member of Velasquez from the great state of New York for her five minutes.
▶ 2:06:04Thank you, Mr. Chairman, and thank you to all the witnesses for being here. Um, and I'm sorry I wasn't when you were uh making your statements because I was participating in another hearing on financial services. Uh, Mr. Chicago, I concur that transparency uh in AI and information on AI is crucial.
▶ 2:06:30Um and for members of this committee and other committees of jurisdiction regarding AI policy, accurate data is important in order for us uh to make uh effective AI policy. Um, estimates for small business AI adoption rates vary widely.
▶ 2:06:52The SBA's office of advocacy estimates the rate is 8.8% but the US chamber estimates that it is at 58% and nearly 50 point uh gap. Why does this different exist? probably a long answer, but candidly it depends on what you how you define adoption.
▶ 2:07:18Uh is adoption an employee who uses chatbt to answer a question once twice a day or is it kind of working truly sidebyside with a number of AI tools to do your work? That's a wide definition. What I would say is when I think about the 200 businesses that we
▶ 2:07:33or even ask any CEO in the private sector currently, no one will say that that AI adoption is where they want it to be. All of them would say that their workforces are not where they want to be to get the full value for their business and for their community. And so the question is how do we increase that that adoption and I think the that goes to the trust questions that goes to the infrastructure questions goes to the literacy questions. But I think right now it depends on how you define adoption. But if you were to ask a CEO what do they want from adoption?
▶ 2:07:58They want I want my people using AI to do the things that people shouldn't be doing and so that they can do the things that only people can do so that our business grows, we make more revenue and we help our communities and grow jobs.
▶ 2:08:07Thank you. and bias and discriminatory decision making are major AI risks. How can this harm small firms and incur legal and reputational costs? I
▶ 2:08:21I'll I'll use a very specific example. This actually comes from my time when I was serving federally. Um but we would hear from small businesses who are using AIdriven uh resume and hiring tools uh to comb rums to begin to kind of sift out people for that first round. And what they began to find, this is particularly became an issue during the pandemic, was at times people who could do the job were unintentionally screened out, which of course for the business created a problem of there's an empty billet of someone who could have done a job um and now that job is not is not being done.
▶ 2:08:50That's one very specific example where not only do we do does that type of bias um offend our values, but it also goes against kind of common sense in the interest of the business. So the key is how do we make sure that um we mitigate bias not just for our values but for the sake of the business who's actually trying to get the benefit that that it's the right thing to do but it's also the smart thing to do.
▶ 2:09:12So can training data be uh adjusted to reduce bias.
▶ 2:09:17Uh the answer is yes but it's the important thing is that that is not on the small business. It can't be they're not making the tool. uh if I this goes to I think frankly the largest folks who are making these tools and the and the key thing that I would say is it's in their interest to do such. If I'm a small business and I'm creating a tool, I want it to reflect my community so I can serve my customers better. Uh tools that don't do that are not going to help me serve my customers.
▶ 2:09:42So if I'm a provider, an LLM provider, a hyperscaler who's trying to create a tool, you want to help your customers help their customers. And having representative uh data that actually reflects reality and the outcomes that it produces is best for everyone. Thank you, Mr. Horn. You write in your testimony, the principle behind the AI wise act that federal support should remain neutral among specific tools and vendors is exactly right. Why?
▶ 2:10:11What are the consequences if the SBA and other agencies favor certain vendors over
▶ 2:10:20Thank you uh ranking member for the question. uh we are in the very early days of AI. It's too soon to start picking winners and not winners. Um in our own work with hundreds of organizations, we see that different tools are right for different purposes and that uh organizations that succeed don't just rely on one single AI tool. They they use different tools. They combine them. They bring humans.
▶ 2:10:43I truly believe the future of of work and teaching and learning are multiple AIs and multiple people working together in a collaborative environment. Thank you. Thank you, Mr. Chairman. Yield back.
▶ 2:10:55Gentle lady yields back. And I'd like to thank our witnesses for their testimony today and for appearing before us. It's been great dialogue. Without objection, members have five legislative days to submit additional materials and written questions for the witnesses to the chair which will be forwarded to the witnesses. I ask the witnesses to please respond promptly. If there's no further business, without objection, the committee is now adjourned.