“Shaping Tomorrow: The Future of Artificial Intelligence”
Hearing•House Oversight and Government Reform Subcommittee on Cybersecurity, Information Technology, and Government Innovation•Sep 17, 2025 · 2:00 PM
Summary
House Oversight and Government Reform Subcommittee on Cybersecurity, Information Technology, and Government Innovation held a hearing on Sep 17, 2025 at 2:00 PM in Rayburn House Office Building, Room 2247. 3 witnesses appeared.
Record
The meeting has its video, its transcript, witnesses and documents on the record.
Video
The proceedings, as the committee streamed them.
Transcript
The transcript runs to 1,794 lines and 93,343 characters, as the Government Publishing Office printed it.
1[House Hearing, 119 Congress]2[From the U.S. Government Publishing Office]34 SHAPING TOMORROW:5 THE FUTURE OF ARTIFICIAL INTELLIGENCE67=======================================================================89 HEARING1011 BEFORE THE1213 SUBCOMMITTEE ON CYBERSECURITY,14 INFORMATION TECHNOLOGY,15 AND GOVERNMENT INNOVATION1617 OF THE1819 COMMITTEE ON OVERSIGHT AND GOVERNMENT REFORM2021 U.S. HOUSE OF REPRESENTATIVES2223 ONE HUNDRED NINETEENTH CONGRESS2425 FIRST SESSION2627 __________2829 SEPTEMBER 17, 20253031 __________3233 Serial No. 119-493435 __________3637Printed for the use of the Committee on Oversight and Government Reform3839[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]4041 Available on: govinfo.gov, oversight.house.gov or docs.house.gov4243 __________4445 U.S. GOVERNMENT PUBLISHING OFFICE4661-734 PDF WASHINGTON : 20254748-----------------------------------------------------------------------------------49 COMMITTEE ON OVERSIGHT AND GOVERNMENT REFORM5051 JAMES COMER, Kentucky, Chairman5253Jim Jordan, Ohio Robert Garcia, California, Ranking54Mike Turner, Ohio Minority Member55Paul Gosar, Arizona Eleanor Holmes Norton, District of56Virginia Foxx, North Carolina Columbia57Glenn Grothman, Wisconsin Stephen F. Lynch, Massachusetts58Michael Cloud, Texas Raja Krishnamoorthi, Illinois59Gary Palmer, Alabama Ro Khanna, California60Clay Higgins, Louisiana Kweisi Mfume, Maryland61Pete Sessions, Texas Shontel Brown, Ohio62Andy Biggs, Arizona Melanie Stansbury, New Mexico63Nancy Mace, South Carolina Maxwell Frost, Florida64Pat Fallon, Texas Summer Lee, Pennsylvania65Byron Donalds, Florida Greg Casar, Texas66Scott Perry, Pennsylvania Jasmine Crockett, Texas67William Timmons, South Carolina Emily Randall, Washington68Tim Burchett, Tennessee Suhas Subramanyam, Virginia69Marjorie Taylor Greene, Georgia Yassamin Ansari, Arizona70Lauren Boebert, Colorado Wesley Bell, Missouri71Anna Paulina Luna, Florida Lateefah Simon, California72Nick Langworthy, New York Dave Min, California73Eric Burlison, Missouri Ayanna Pressley, Massachusetts74Eli Crane, Arizona Rashida Tlaib, Michigan75Brian Jack, Georgia Vacancy76John McGuire, Virginia77Brandon Gill, Texas7879 ------8081 Mark Marin, Staff Director82 James Rust, Deputy Staff Director83 Mitch Benzine, General Counsel84 Lauren Lombardo, Deputy Policy Director85 Raj Bharwani, Senior Professional Staff Member86 Duncan Wright, Senior Professional Staff Member87 Mallory Cogar, Deputy Director of Operations and Chief Clerk8889 Contact Number: 202-225-50749091 Robert Edmonson, Minority Staff Director92 Contact Number: 202-225-505193 ------9495 Subcommittee on Cybersecurity, Information Technology, and Government96 Innovation9798 Nancy Mace, South Carolina, Chairwoman99100Lauren Boebert, Colorado Shontel Brown, Ohio, Ranking101Anna Paulina Luna, Florida Member102Eric Burlison, Missouri Ro Khanna, California103Eli Crane, Arizona Suhas Subramanyam, Virginia104John McGuire, Virginia Yassamin Ansari, Arizona105106 C O N T E N T S107108 ----------109110 OPENING STATEMENTS111112 Page113114Hon. Nancy Mace, U.S. Representative, Chairwoman................. 1115116Hon. Shontel Brown, U.S. Representative, Ranking Member.......... 2117118 WITNESSES119120Ms. Kinsey Fabrizio, President, Consumer Technology Association121Oral Statement................................................... 4122123Mr. Samuel Hammond, Chief Economist, Foundation for American124 Innovation125Oral Statement................................................... 5126127Dr. Nicol Turner Lee (Minority Witness), Senior Fellow,128 Governance Studies, Director, Center for Technology Innovation,129 The Brookings Institution130Oral Statement................................................... 7131132Written opening statements and bios are available on the U.S.133 House of Representatives Document Repository at:134 docs.house.gov.135136 INDEX OF DOCUMENTS137138 * Article, Google, ``AlphaEvolve, A Gemini Powered Coding Agent139 for Designing Advanced Algorithms''; submitted by Rep.140 Burlison.141142The documents listed above are available at: docs.house.gov.143144 ADDITIONAL DOCUMENTS145146 * Questions for the Record: Dr. Nicol Turner Lee; submitted by147 Rep. Brown.148149These documents were submitted after the hearing, and may be150 available upon request.151152 SHAPING TOMORROW:153 THE FUTURE OF ARTIFICIAL INTELLIGENCE154155 ----------156157 WEDNESDAY, SEPTEMBER 17, 2025158159 U.S. House of Representatives160161 Committee on Oversight and Government Reform162163 Subcommittee on Cybersecurity, Information Technology, and Government164 Innovation165166 Washington, D.C.167168 The Subcommittee met, pursuant to notice, at 2:02 p.m., in169room 2247, Rayburn House Office Building, Hon. Nancy Mace170[Chairwoman of the Subcommittee] presiding.171 Present: Representatives Mace, Burlison, Crane, McGuire,172Brown, and Subramanyam.173 Ms. Mace. Good afternoon. The Subcommittee on Cybersecurity174Information, Technology, and Government Innovation will now175come to order, and welcome everyone.176 Without objection, the Chair may declare a recess at any177time. I recognize myself for the purpose of making an opening178statement.179180 OPENING STATEMENT OF CHAIRWOMAN NANCY MACE181182 REPRESENTATIVE FROM SOUTH CAROLINA183184 Good afternoon and thank you all for being here for today's185important hearing on the future of artificial intelligence.186From the tools powering your smartphone to the algorithms187predicting weather, recommending medicines, or helping farmers188improve crop yields, AI is already shaping the world around us.189 Just as we once competed for dominance in space or nuclear190technology, the United States is now in a race for leadership191in AI. American companies are at the frontier for this race.192These companies are pushing the boundaries of what advanced193language models can do, and countless startups and research194labs are finding new applications for AI in every corner of the195economy.196 The stakes are high. If the United States leads, we get to197shape the standards, the ethics, and the economic benefits of198this powerful technology. If we fail, we cede such influence to199adversaries who do not share our values. So, the risks are200high.201 AI will have an impact on all Americans across all202industries. AI is driving new efficiencies and creating203breakthroughs to improve lives. In healthcare, AI is helping to204detect cancer early and accelerating drug development. And205transportation is making cars safer and logistics smarter. In206agriculture, it is reducing waste and helping farmers feed more207people with fewer resources.208 These advances are not abstract. They are happening now and209are creating better services, lower costs, and new210opportunities for American workers and American families. But211the technological future of AI remains uncertain.212 Some experts warn we are just a few years away from the213emergence of artificial general intelligence or the214singularity. Others argue that technology has inherent215limitations, and we are decades away from the singularity, if216it is even possible.217 We do not know for certain what future of AI will look218like, but what I do know is the future is too important to219leave up to chance. We are going to do our best to understand220what kinds of impact AI can have on our economy, our society,221and develop potential solutions now before it is too late.222 This Subcommittee takes seriously its responsibility to223examine these issues, and I am looking forward to hearing today224from everyone on both the current state of AI and the possible225futures which lie ahead.226 It is essential the United States lead, not just in227building these technologies but ensuring they are developed228responsibly, deployed safely, and used in ways which advance229American values. When we get this right, we will ensure230artificial intelligence fulfills its extraordinary promise.231 I look forward to today's discussion and to working with my232colleagues on this Committee to ensure America leads in shaping233the future of AI.234 I now recognize Ranking Member Brown for her opening235statement.236237 OPENING STATEMENT OF RANKING MEMBER238239 SHONTEL BROWN, REPRESENTATIVE FROM OHIO240241 Ms. Brown. Thank you, Chairwoman Mace.242 Artificial intelligence is here, and it is already243reshaping our economy, workforce, and daily life. As we work to244ensure that America leads in AI innovation, we must also lead245in responsible and trustworthy use of this technology.246 AI holds the promise to strengthen our economy and make247government more efficient. However, when commonsense safeguards248are absent, technology can deepen inequalities, leave workers249behind, or allow bad actors to take advantage of gaps in250policy.251 Even as we look toward the future, we cannot ignore the252ways AI is already changing the workplace. And, while some of253these changes are promising, we must also work to prepare the254American people for change.255 Workers in my Cleveland district and across the country are256worried about what automation and emerging technologies mean257for their job and their security. Black workers in particular258remain disproportionately concentrated in positions most at259risk of automation according to research by McKinsey & Company.260If we fail to provide retraining, education, and pathways into261the jobs of the future, we risk leaving entire communities262behind.263 A diverse prepared workforce is not just good for our264economy, but a necessity for our national competitiveness. If265we do not ensure that employees most at risk of being replaced266by AI have other pathways for employment, adoption of AI will267not only drive greater economic disparity, it will also miss268opportunities to diversify and elevate the workforce. A diverse269and adequate workforce not only builds up our communities, it270also advances our AI ambitions.271 We know that foreign adversaries, particularly the Chinese272Communist Party, are aggressively pursuing technological273dominance. They are not only racing to outpace us in artificial274intelligence and cybersecurity, but also actively targeting our275institutions, businesses, and citizens.276 At the same time, everyday Americans face scams, fraud, and277data breaches that threaten their livelihoods and erode trust278in government and the private sector. The future of AI will be279shaped by our commitment to getting it right today and our280ability to learn serious lessons and mitigate future risk.281 That is why our oversight work must focus on several areas:282defending against hostile foreign governments, holding283accountable scammers who prey on vulnerable communities, and284investing in our workforce to ensure resiliency. We must ensure285that innovation does not come at the expense of fairness,286security, or opportunity.287 Thank you to all the witnesses that are here today to288discuss this critically important topic.289 And, with that, I yield back.290 Ms. Mace. Thank you, Congresswoman Brown.291 I am pleased to introduce our witnesses for today's292hearing.293 Our first witness today is Ms. Kinsey Fabrizio, President294of the Consumer Technology Association.295 Our second witness is Mr. Samuel Hammond, Chief Economist296at the Foundation for American Innovation.297 And our third witness today is Dr. Nicol Turner Lee, Senior298Fellow of Governance Studies and Director of the Center for299Technology Innovation at the Brookings Institute.300 Welcome everyone, and we are pleased to have you this301afternoon. And pursuant to Committee Rule 9(g), the witnesses302will please stand and raise your right hand.303 Do you solemnly swear or affirm that the testimony that you304are about to give is the truth, the whole truth, and nothing305but the truth, so help you God?306 Let the record show that the witnesses all answered in the307affirmative.308 We appreciate all of you being here today. You may sit back309down, and we look forward to your testimony.310 Let me remind the witnesses that we have read your written311statements, and they will appear in full in the hearing record.312Please limit your oral statements to 5 minutes.313 As a reminder, please press the button on the microphone in314front of you so that is on and the Members can hear you. When315you begin to speak, the light in front of you will turn green.316After 4 minutes, the light will turn yellow. When the red light317comes on, your 5 minutes has expired, and we would ask that you318please wrap it up.319 So, we will first recognize Ms. Fabrizio to please begin320her opening statement.321322 STATEMENT OF KINSEY FABRIZIO, PRESIDENT323324 CONSUMER TECHNOLOGY ASSOCIATION325326 Ms. Fabrizio. Thank you. Good afternoon, Chairwoman Mace,327Ranking Member Brown, and Members of the Subcommittee. Thank328you for holding this hearing and for the opportunity to329testify.330 CTA represents over 1,200 companies, and more than 80331percent are startups, small, and mid-sized businesses. Our332numbers power the American economy and support more than 18333million jobs. But, before I talk about AI's broader impact, I334want to share how it is reshaping my own life.335 I am a wife, and I am a mom of two wonderful kids, and I336also run CES, the world's most powerful technology event. Each337week, I use AI to organize my life at home, pulling together338school pickups, drop-offs, sports schedules; and, at work,339these tools help me make smarter decisions, research340competitors, and come up with new ideas for products and341services.342 In many ways, AI is a personal assistant that gives me back343time, time I can spend with my family while staying focused on344leading a complex organization, and it makes a big difference345in my daily life. I am really excited to share the even greater346impact it has at scale across society.347 While so much of the public debate around AI focuses on how348this technology might evolve in the future, AI is here now, and349it is integrated into our lives and delivering benefits for350millions of Americans. We see these technologies in action at351CES, where innovators come together, from AI-powered health352insights from Abbott and Withings, to John Deere's autonomous353tractor, to Oshkosh and Waymo's collision avoidance and354autonomous technologies, and even Siemens' digital twin355platform for manufacturing.356 These products are already in the market and making amazing357changes. Still, we are just scratching the surface of what AI358can do.359 Today we see AI and digital twins that can simulate360everything from factories to city planning; agentic AI, which361are autonomous systems that can manage everyday tasks; vertical362AI models, which are specialized in areas like healthcare and363mobility or agriculture; industrial AI, which is augmenting the364workforce and improving safety; and physical AI, which includes365more lifelike and useful robots.366 American companies are leading the AI race, but their367success is not guaranteed. In China, the government has made AI368central to its national strategy and invested heavily in areas369like semiconductors, robots, and data centers.370 To counter this strategy, we need policies that help371American companies out-innovate the competition. If America372falters in AI, we risk ceding entire industries, supply chains,373and influence over global standards.374 That is why CTA has urged Congress to adopt a 10-year pause375on enforcement of state and local AI laws. In 2025 alone,376legislators across all 50 states introduced more than a377thousand often conflicting AI-related bills. For a startup or a378small business, navigating this patchwork is crippling.379 A pause gives Congress the time it needs to develop a380preemptive Federal framework for AI. The Administration's381recently released AI Action Plan is a powerful and positive382blueprint ensuring American AI innovators have the guardrails383they need to build, grow, and compete.384 We also need a comprehensive Federal privacy law to power385up innovation with more clarity and protect consumers and lower386compliance costs for industries that rely on responsible data387use and give Americans confidence in these life-changing388technologies.389 Congress must also recognize where our laws and frameworks390are working. The law is clear. Simply reading or processing391content does not constitute infringement. This clarity is a392huge competitive advantage for America, and it is the393foundation that allows U.S. companies from the smallest startup394to the largest global brand to win the AI race.395 America has led every major technological wave from396electricity to the internet, and if we get AI policy right,397this technology will be the next great American growth engine.398If we get it wrong through fragmented or restrictive399regulation, like the EU's AI Act, we risk exporting those jobs400and that leadership overseas.401 CTA believes the path is clear: foster innovation, protect402consumers, and ensure America sets the rules of the road for403AI. I look forward to working with this Committee on a404bipartisan basis to shape our AI future.405 Thank you, and I look forward to your questions.406 Ms. Mace. Great. I now recognize Mr. Hammond to please407begin his opening statement.408409 STATEMENT OF SAMUEL HAMMOND, CHIEF ECONOMIST FOUNDATION FOR410 AMERICAN INNOVATION411412 Mr. Hammond. Chairwoman Mace, Ranking Member Brown, Members413of the Committee, thank you for the opportunity to testify414today.415 My name is Samuel Hammond. I am the Chief Economist for the416Foundation for American Innovation. We are a group of417technologists and policy experts focused on developing418technology, talent, and ideas to support a freer and more419abundant future.420 The capabilities of frontier AI systems are improving at a421stunning rate. Five years ago, large language models could422barely generate coherent English text. Today, they can hold423forth on any topic, reason through Ph.D.-level math problems,424and code entire applications from scratch.425 Recent AI progress, including the rise of reasoning models426and AI agents, has been largely driven by breakthroughs in427reinforcement learning applied to Large Language models (LLM)s.428Language models gain their raw intelligence by predicting429sequences of text but, with reinforcement learning, can be430trained to follow instructions, use tools, and pursue complex431goals.432 Scalable reinforcement learning for language models to give433them reasoning goal-directed behavior was only unveiled a year434ago but is already driving rapid improvements in domains like435math and programming. The scope and significance of this436breakthrough is still not fully appreciated, though.437 In principle, these techniques can be used to create438superhuman AI agents in any domain where success can be439objectively benchmarked. Math and software engineering are just440the low-hanging fruit.441 The AI research organization Model Evaluation & Threat442Research (METR) carefully measures progress in AI autonomy and443has found that the length of tasks that AI agents can perform444doubles roughly every four to seven months, a trend that has445held for the past six years.446 While the earliest chat bots could only perform tasks447measured in seconds or minutes, OpenAI's latest model, GPT-5,448can coherently execute tasks that take human engineers two449hours and 17 minutes on average. If this trend continues, we450are only two doublings away, roughly eight to 14 months, from451AI agents that can autonomously perform tasks that take humans452a full 8-hour workday.453 Progress in nonverifiable and open-ended domains is also454accelerating, recently leading to some of the first major455examples of AIs that have made novel scientific and456mathematical discoveries. It is now plausible that we will have457the first superintelligent AI scientists and mathematicians by458the year's end, portending a dramatic speed-up in the pace of459R&D going forward. This includes AIs optimized for AI research460itself, creating the glimmers of a self-improving feedback loop461whereby AIs rapidly help build their own successors.462 What happens when AIs get better at AI R&D than the best463human researchers in the world? At a minimum, we should expect464a discontinuous leap in the power and efficiency of the465frontier models. But where this process tops out is still a466matter of significant uncertainty.467 It is possible that, even with fully automated AI R&D,468progress will remain bottlenecked by the availability of469compute, data, and energy. It is also possible that we are only470one or two major breakthroughs away from systems that can learn471continuously in an unbounded fashion.472 Regardless, the jump in capabilities unlocked by473recursively self-improving AI is likely to be profound, even474within the bounds of existing infrastructure, and is coming475sooner than many realize.476 It is worth emphasizing that creating Artificial General477Intelligence (AGI) and superintelligent AI that is capable of478outperforming humans in every domain is the explicit goal of479every leading U.S. AI company. While some dismiss this as480science fiction or marketing hype, I assure you the leaders of481these companies are deadly serious.482 As for timing, Anthropic cofounder Jack Clark testified483recently that he expects transformative AI to arrive as soon as484the end of 2026 or early 2027. Even if these forecasts are on485trend, AI capabilities will remain uneven for at least several486more years.487 For a brief paradoxical moment, we will have488superintelligent AIs that can prove new math theorems but still489struggle to do many things that humans find trivial. This is490especially true in areas like robotics, which, despite491remarkable progress, are still many years away from492outperforming humans in every physical domain given the paucity493of high-quality training data.494 So, as we run headlong into this new world, I see four495major takeaways for national policymakers.496 First, monitoring frontier AI capabilities in real time497should be a national security imperative of the U.S.498Government. Early and differential access to the developments499of the frontier can provide policymakers and national security500advisers with the foresight into the capabilities that are501coming down the pike, giving us time to prepare and adapt.502 Second, as AI systems become human-level and beyond,503geopolitical power will be increasingly proxied by the global504distribution of computing resources. America's existing lead is505downstream of our massive advantages in AI hardware and data506centers, but this is tenuous at best. With China outbuilding us507on new energy, we must double down on semiconductor export508controls or risk being leapfrogged.509 Third, we must quickly advance the frontier of AI control510interpretability, review our laws and regulations for the511compatibility with powerful AI, and invest in much more robust512cyber and infrastructure security, all priorities outlined in513President Trump's AI Action Plan.514 Fourth and finally, we must open our minds to radically new515forms of institutions and structures of government. From the516printing press to the Industrial Revolution, every major517technological transition has driven equally transformative518changes to our system of government. I believe the AI519revolution will be no different.520 It raises unique challenges given AI's use cases for521surveillance and censorship, as seen in China's model of the522digital panopticon. Reconciling the advent of powerful AI523systems with America's tradition of individual liberty and524limited government is, thus, the challenge of our time.525 Thank you, and I look forward to your questions.526 Ms. Mace. Thank you.527 I now would like to recognize Dr. Turner Lee for 5 minutes.528529 STATEMENT OF NICOL TURNER LEE (MINORITY WITNESS)530531 SENIOR FELLOW, GOVERNANCE STUDIES532533 DIRECTOR, CENTER FOR TECHNOLOGY INNOVATION534535 THE BROOKINGS INSTITUTION536537 Dr. Turner Lee. Thank you. Thank you, Chairwoman Mace,538Ranking Member Brown, and distinguished Members of the539Committee for this invitation to testify.540 My research focuses on policies that govern AI, digital541divide, as well as innovation. Artificial intelligence is not542the future; it is here. Today several workplaces require use of543AI by workers, and one report shows that 92 percent of544companies have plans to increase their investment in the545technology.546 In just about every sector, companies are figuring out the547role of AI-enhancing productivity, as well as the most548appropriate investment in talent. Beyond enterprise use cases,549AI is transforming the delivery of critical services, domains,550such as government services, healthcare, and education. And,551with so much on the horizon, it is imperative that we strike552the correct balance of innovation and regulation.553 We must safeguard consumers, institutions, and critical554infrastructure from AI risk, including workplace displacement,555bias and discrimination, and the irreparable harm of machines.556 In my testimony today, I just want to offer three points557that I think Congress should take, which will be critical to558the future of AI: the need for responsible and ethical559frameworks of AI design and governance, the importance of a560ready and agile talent pipeline and workforce, and the561importance of monitoring the unknowns in AI to ensure its562safety and security while addressing the clear and present563challenges.564 AI depends on responsible and ethically designed models in565national governance. The stakes are too high. The rapid566advancements of generative AI in video text and voice567extraction have contributed to consumer fraud. Our Nation's568seniors are increasingly being targeted, in some cases falling569for financial AI voice cloning and deepfake scams that ask them570to send their money to relatives.571 Being responsible requires national governance, and we have572laws, existing laws, for highly regulated industries that573protects consumers in digital spaces, including the Fair574Housing Act, the Equal Credit Opportunity Act, among others.575But, without congressional resolution in other critical policy576areas, like data privacy, everyday Americans will be exploited577by malicious uses of AI systems.578 We need clear measures that ensure human oversight,579disclosures, and independent audits over automated and580autonomous decisions. We have started this process with the AI581Action Plan this past summer, but our focus on promoting U.S.582leadership against China as the prime goal will only allow583these consumer protection goals to fall to the wayside.584 Yes, recently, states, in the absence of Federal585legislation, have actually moved forward with their own586legislation. And, since January, over a hundred measures across58738 states have been enacted to law. Multiple state attorney588generals have also issued guidance on how to apply these laws589to AI.590 The rejected proposed 10-year moratorium on states would591have threatened states' rights and the public interest, and592leadership needs to continue to protect the independence of593Federal agencies so they can serve as the bulwark against594deceptive and unfair consumer-facing AI applications. Cuts to595these agencies weaken our ability to hold bad actors596accountable, but they also undermine consumer trust.597 Second, the future of AI depends on a ready and agile598talent pipeline and workforce. Maintaining our edge means that599we have robust talent that incorporate diverse viewpoints in600the design, development, and deployment of AI.601 Immigrants are central to the story of innovation. 77602percent of the top AI companies were founded or cofounded by603first-generation immigrants. Policies that restrict immigration604may threaten our innovation capabilities, and defunding605research at world-class universities and scientific606institutions may also do the same.607 Let me just share a few statistics of what is happening as608a result of other countries exploiting these opportunities.609Spain opened its doors to students subject to U.S.610restrictions. European universities are offering scientific611asylum to scientists. China is using the reverse brain drain to612aggressively recruit our top technological talent.613 Cultivating talent at that level has impacts as well as614ensuring that researchers here in the United States have what615they need to get the research done for the next big idea, and616we cannot neglect our domestic talent pipeline. Recent declines617in math and reading scores are warning signs that much more618needs to be done to cultivate homegrown talent, not only619through national apprenticeships, but also realigning our core620objectives in schools so that we meet the demands of the future621workforce.622 Casualties will abound in the workplace if we have a less623agile, ready workforce capable of shifting gears.624 And, finally, I will just say this. Though the future of AI625is largely unknown, we need to solve the first point I made and626the second point to ensure that we actually get to a place627where we understand the power of artificial intelligence or628artificial general intelligence, agentic AI, as well as what is629largely unknown at this point.630 However, people like me are far more skeptical of AGI and631generally the existential threats for the foreseeable future632simply because we just do not know enough, and it is important633for us to have those guardrails in place so that we ensure the634rights and safety of all Americans central to that development,635and we not abandon consumer protection in the rush to just636innovate.637 So, I will close here and implore Congress to continue to638think about policies that allow us to grow a healthy ecosystem639where consumers are centered and our economy and our trust in640these products are also prioritized.641 Thank you again to the Members of the Committee. I look642forward to working with you and taking your questions.643 Ms. Mace. Thank you, ma'am.644 And I will now recognize myself for 5 minutes and for645questioning.646 Ms. Fabrizio, thank you for being here today. What has647surprised you the most about AI?648 Ms. Fabrizio. At CES, I think the most surprising thing are649all the different ways that AI is solving the world's biggest650challenge. The healthcare applications are the most exciting to651me personally. I saw a digital twin of a heart at CES, which is652used to train surgeons so that they can understand how to653safely do heart surgery. That is a huge impact and very654amazing.655 Ms. Mace. And then you talked about the 10-year moratorium656and for states. Why is it so important--the states' rights657thing has to be balanced, but also we do not want to stifle658innovation.659 We know China, Russia, Iran, they are not--they do not have660any guardrails, they do not care. Talk about that a little bit.661 Ms. Fabrizio. Yes. Well, you said it. China does not have662that. It is impossible for our member companies--like I said,663we have 80 percent small businesses, and they cannot compete664and understand when there are a thousand different potential665laws that they have to comply to. It just stifles innovation666completely. And, for us to win the AI race, we need to remove667that barrier.668 Ms. Mace. And it is a Federal issue because it is commerce669across state lines, and having all those--that regulatory670environment--a patchwork in every state--does make it very671difficult to operate.672 And, then, Mr. Hammond, one thing that struck me in your673testimony, you talk about compute energy. Talk to me a little674bit more about that.675 Let us go into detail because I agree with you. It is a676huge problem. How do we solve it?677 Mr. Hammond. It is a great question. So, you know, there678are only a handful of inputs that go into training and679competing at the frontier with these models. There is the data,680the human talent, the compute, and the energy.681 With China, we are basically at parity with talent. With682data, they may have advantages because they do not have privacy683laws. They can----684 Ms. Mace. They have stolen a bunch of our data, right?685 Mr. Hammond. Of course. And they also steal data and686intellectual property (IP).687 Ms. Mace. Yes.688 Mr. Hammond. And so, really, it comes down to hardware689energy. China has added over 400 gigawatts to their grid last690year. They are about to do the same thing this year, so only--691--692 Ms. Mace. How much have we added to our grid?693 Mr. Hammond. Approximately zero. I mean, we have removed694coal and added renewables. And that has canceled out.695 So, what that means is, in lieu--but for these export696controls that are barring China from our most advanced697hardware, they would surely leapfrog us within a matter of698years.699 Ms. Mace. You think nuclear is the way?700 Mr. Hammond. I support nuclear. I think the earliest that701we will see new reactors come online is in the 2030s.702 Ms. Mace. You know, it is frustrating because we see small703nuclear reactors, or SMRs, in Japan and in France, and we do704not have them here. We have them in our nuclear subs.705 Like, I just--can we just--it is a joke, but it is like why706can we not just plug one into an outlet? I mean, I just--we707have the technology here. Why are we not using it--particularly708with our data centers--and allow them to grow as data center709technology--as data centers to grow as well.710 But I do want to talk to you about the future of AI. Elon711Musk has said that, as early as 2026, we would have712singularity, basically. Define singularity, and how quickly do713you think we are going to get there?714 Mr. Hammond. So, I think there are two ways to think of715singularity. One is as a metaphor for our ability to predict716the future. So, whatever the technology it is, a singularity is717a point in time where it becomes impossible to look beyond that718point.719 And then there is the technological singularity, which is,720when we have AIs, they can build their own successors, and then721potentially go off to infinity, and we do not really know what722comes out of that process.723 I think 2026 is a pretty aggressive expectation, but I724think something like that crossing that threshold will happen725this decade.726 Ms. Mace. And I have talked with some folks in tech space727that say a thousand days or 2,000 days.728 What do you think it will take to get there, where AI is729creating its own AI?730 Mr. Hammond. So, you know, I got into this topic as a young731kid reading Ray Kurzweil's 1999 book where he predicted we732would have human-level AI by 2029. That was a 30-year forecast,733and actually current trend lines suggest he was dead on 30734years ago. And so, I tend to lean toward that as a date.735 But I think this will not look like some threshold that we736pass and looks completely different. I think we are on this,737sort of, continuous exponential. And the thing about738exponentials is they look flat looking backward and vertical739going forwards.740 Ms. Mace. What is the biggest concern with when we do get741there, when we do hit that milestone?742 Mr. Hammond. That we lose control in some manner, whether743literally over the AIs themselves or that--the proliferation744because, while it does cost billions of dollars to train these745models at first, subsequent generations, the cost comes down in746orders of magnitude. That there will be a mass proliferation of747powerful capabilities that our institutions are just not748capable of adapting to.749 Ms. Mace. That can hack every system, every grid everywhere750all at the same time all around the world, essentially,751potentially. That is the way I see it.752 Mr. Hammond. Yes. I think there are going to be attacks on753critical infrastructure, but there will also be, you know, the754high school student that hacks their school's IT system and the755system admin happens to be the gym teacher.756 Ms. Mace. Or personal medical records or whatever the case757may be. Yes, that is one of my greatest concerns on the cyber758side.759 I have run out of time. I could have--I could talk all day.760Maybe we will have some more time later.761 So, I will yield back to the Ranking Member and recognize762her for 5 minutes.763 Ms. Brown. Thank you, Madam Chair.764 Artificial intelligence has the potential to be in every765aspect of the workforce in every corner of our daily life. This766means that the Federal Government and private sector must767collaborate to adapt to an AI present and future.768 Algorithms and automation are not inherently harmful, but769the way in which they are developed and deployed has the770potential to have profound consequences for American workers,771especially those from diverse backgrounds.772 Black and Brown communities have long carried the weight of773the wage gap in this country, and we cannot allow AI to deepen774those inequities, whether through bias algorithms in hiring or775automation that displaces jobs. We cannot afford to be caught776flatfooted or let AI run unchecked.777 So, Dr. Turner Lee, can you speak of the work that AI778Equity Lab and Brookings is doing to ensure that AI does not779worsen historical inequities and the importance of acting780proactively?781 Dr. Turner Lee. Thank you, Ranking Member.782 So, what I actually figured out as a sociologist is that we783are not a lot of sociologists sitting at the table or people of784various backgrounds when it came to thinking about the outcomes785of these models. And so, writing on the back of a napkin, I786thought about an experiment to actually bring into disciplinary787experts, people from various backgrounds, various industries788together to think about areas in which we are going to have the789most high risk and consequential outcomes, particularly with790marginalized communities.791 To your point, not only do some AI models come with a792series of bias in the training data where it is actually793picking up information that may be false, inaccurate, or under-794representative, the outcomes of that data can contribute to a795widening wealth gap, when that algorithm suggests that I am not796creditworthy, I cannot get a home loan, in essence, my quality797of life is actually impacted.798 So, I think one of the best ways to start with this is to799widen the seats at the table and to ensure that we have800scientists, alongside social scientists, alongside industry801sector, alongside people with various backgrounds who have802concerns with the lived experiences of populations that you803spoke about, especially those that are Black and Brown804communities.805 Ms. Brown. Thank you. And, as more AI companies adopt AI806technology, we must accept that the future of work is changing.807 Dr. Turner Lee, what steps can the Federal Government take808to ensure that American workforce is prepared to succeed in the809AI future, and what legislative steps can Congress take to810ensure that there are adequate guardrails overseeing AI811adoption throughout the economy and society while also812encouraging AI innovation?813 Dr. Turner Lee. Well, first and foremost, I would like to814commend the work you have done in a bipartisan manner on your815act, which is around training the Federal workforce to be816exposed to AI. I think that is the first step.817 But, to be very transparent, again, the train has left the818station. AI is not only dictating how they do work, but it is819also managing their productivity in how they are processing820that work as well. So, it may help them with research, but it821is managing the time that it takes to do the research.822 Being transparent with the Federal Government, I think, is823one way to actually help disclose the use of AI there, also824making available all types of data. We have some concerns,825given the recent scrub of a lot of information from Federal826datasets, et cetera, that there just will not be the quality of827data and integrity of data that we need. So, just making sure828the Federal Government stays on point with that and does the829appropriate audits of the data that is available.830 I would just suggest, in terms of guardrails, there is831enough AI for everybody to eat is a statement that I have been832making lately. And that means that, wherever you are in the833workforce, you are going to in some way touch this. Improving834upon our literacy, our upskilling, our ability to mentor people835who may be from different generations where AI was not836necessarily something that they ever thought of--I used to837watch the Jetsons. I never thought that AI would come to838fruition. I think, Ranking Member, those are the steps that I839think the U.S. Government can put in place that really in the840long run promote transparency, disclosure, and effective use.841 Ms. Brown. Thank you very much. I will close with this. Our842future is one where AI technology will, no doubt, impact our843everyday lives, which is why we must carefully consider AI844development now. We need Federal legislation that protects845Americans' rights and freedoms by preventing bias and846injustice.847 I look forward to continuing to explore this topic with848experts so that we can ensure safe and responsible innovation.849 And, with that, Madam Chair, I yield back.850 Ms. Mace. Thank you.851 I will now recognize Mr. McGuire for 5 minutes.852 Mr. McGuire. Thank you, Madam Chair, and thank you to our853witnesses for being here today.854 For the first few questions, if you would, let us answer as855quickly as possible. Let us see.856 Mr. Hammond, you said you think singularity, based on your857experience, will be 2026. And, just real quick, Dr. Lee, when858do you think we will have that? Just real simple.859 Dr. Turner Lee. I am a little less optimistic. I think it860is going to take longer because I still think it is a little861bit more hypothetical in its framing.862 Mr. McGuire. All right.863 Ms. Fabrizio.864 Ms. Fabrizio. I feel the same way. I think right now human865in the loop is still important, and we are seeing AI augment866what humans can do.867 Mr. McGuire. All right. This is just a yes or no. Do you868believe China is using AI to manipulate their people, or do869they have plans for that? Yes or no, Ms. Fabrizio?870 Ms. Fabrizio. I think China is using AI in ways that we are871not, and that is why it is important for us to continue to872focus on winning the AI race with the issues I laid out873earlier.874 Mr. McGuire. Mr. Hammond?875 Mr. Hammond. Yes. Absolutely, in some cases using U.S.876technologies provided by U.S. companies.877 Dr. Turner Lee. I do agree that the Chinese Government has878a highly surveilled state, and they are using AI not to the879protection of their citizens.880 Mr. McGuire. All right. Just yes or no, because we are881running out of time. Yes or no, the stakes are very high for AI882development in the United States?883 Ms. Fabrizio. Yes.884 Mr. Hammond. Extremely.885 Dr. Turner Lee. Yes.886 Mr. McGuire. So, very important that we have the best887workforce possible to win this battle, yes or no?888 Ms. Fabrizio. Yes.889 Dr. Turner Lee. Yes.890 Mr. Hammond. Yes. In the short run, I wonder what the891workforce will look like when AIs can do everything.892 Mr. McGuire. Yes, I am with you. So, we are developing this893workforce, and I got to tell you, as a Navy SEAL veteran, if894somebody saved my life on the battlefield, I do not care if895they are pink or blue, male or female, Democrat, Republican; we896are all human beings. But I believe that the decisions that we897make should be colorblind because we need the best force, and I898do not care if you are pink or blue.899 Do you agree that it should be colorblind, that we should900have the most qualified people to win this race? Ms. Fabrizio,901yes or no?902 Ms. Fabrizio. We need the best and brightest.903 Mr. Hammond. Colorblind, yes. Not necessarily nationality904blind. One of the challenges here is some of the most sensitive905technologies are being developed by foreign nationals,906including Chinese nationals.907 Mr. McGuire. But, again, if they are the best, they are the908best.909 Mr. Hammond. They are the best of the best. But we have to910compartmentalize in some cases.911 Dr. Turner Lee. I think we should have the best of the912best, but I also think that we need to have the doors open for913people who are the best of the best and all of our communities914to actually participate.915 Mr. McGuire. Okay. Mr. Hammond, I have got a lot to learn,916but I am listening, and I am learning. And I liked what you917talked about, the digital twin and practicing the heart918surgeries. There is so much more I want to learn.919 But, in your testimony, you said, even if we were to reach920AI singularity, we might not have enough energy to keep going.921So, this question is for all of the witnesses, and hopefully it922is an easy one. Let us keep it real simple because we do not923have time.924 Would unleashing American energy give us a better chance of925winning the AI race? And I am talking the ability to drill,926nuclear, all of the above, coal plants.927 Ms. Fabrizio. We need to modernize our energy grid. It928cannot handle what is in store with AI.929 Mr. Hammond. Might be the single-most important factor,930yes.931 Mr. McGuire. Ms. Lee?932 Dr. Turner Lee. I do agree that we need to do more to933upgrade the energy grid, particularly if we are actually934building data centers, but I want us to be cautious about the935environmental consequences of actually moving too fast in936communities where we already know we have a disadvantage.937 Mr. McGuire. And, personally, I would not care about that.938I want to win.939 All right. Ms. Fabrizio, I have to say I agree with your940testimony that overly strict regulations can stifle innovation.941 The first thing I think about, the so-called Green New942Deal, better known as the ``green new scam'', the Biden943Administration spent hundreds of billions of dollars on these944green energy projects, like solar panels and windows.945 Yesterday, I asked an AI chat box, how many acres of solar946panels would you need to power AI in the United States by 2030?947Anyone take a guess how much that would be? A thousand acres.948Actually, let me see here. It is way more than a thousand949acres. It is 500,000 acres. That is half the size of Delaware.950 We should be investing in fossil fuels and nuclear, small951modular nuclear reactors, as we discussed earlier. We will only952need 500 acres to do the same job nationally by 2030. We should953be using coal, natural gas, traditional nuclear power until954SMRs are ready, not solar panels.955 All right. So, let me ask this question. Is China building956thousands of solar farms to power their AI, yes or no?957 Ms. Fabrizio. China is looking at energy in different ways958than we are, but there are solutions that we can look at, too,959to modernize our energy grid, and AI will help--AI will help960develop solutions and help us be smarter about the future.961 Mr. McGuire. Mr. Hammond?962 Mr. Hammond. Both. AI--over 30 new coal plants, while also963adding 300 gigawatts of renewables.964 Mr. McGuire. So, with today's technology, what is more965effective, solar or fossil fuels? Just keep it simple because I966am running out of time.967 Ms. Fabrizio. That--I would have to get back to you on968that.969 Mr. Hammond. I support all of the above. In the short run,970these data centers are only going to go up with natural gas.971 Dr. Turner Lee. I cannot answer the particulars. I can get972back to you that from my team.973 But I want to go back, Congressman, to your question----974 Mr. McGuire. I am very limited on time.975 Dr. Turner Lee. No problem.976 Mr. McGuire. I am sorry. All right. So, let us see.977 Mr. Hammond and Ms. Fabrizio, what are some of ways AI978superintelligence might actually help us solve the energy979problem?980 Ms. Fabrizio. Well, when it comes to research and data and981looking for solutions, AI is faster and can help predict982different models and find different solutions where we may not983be able to find them on our own.984 Mr. Hammond. I believe we are a year or two away from985having fully autonomous AI labs that could discover new986materials, new energy sources, all of the above.987 Mr. McGuire. I am out of time. I yield back. Thank you.988 Ms. Mace. Thank you. I will now recognize Mr. Subramanyam989for 5 minutes.990 Mr. Subramanyam. Thank you, Madam Chair.991 Thank you to the witnesses for coming today.992 I wanted to talk about a couple things. We have had several993hearings on Capitol Hill about AI in recent months and since I994have been here, certainly. And one thing that has not been995talked about as much is job displacement as a result of AI.996 I think--you know, I served in the Obama White House as a997technology policy adviser, and we were talking about this, but998it was a little bit theoretical. I mean, there was job999displacement happening because of emerging technologies,1000certainly, making jobs easier, but automating some jobs and1001some tasks. But now we are seeing it at a different level.1002 We are seeing companies now, basically, lay off entire1003departments and replace them with AI. They are saying this1004publicly, and they are saying this, you know, very--they are1005not hiding it anymore, right? And it is their prerogative. I1006mean, we are not here to tell companies how to run their1007business.1008 But it is creating a problem that Congress has to figure1009out how to address, which is jobs, because one thing we want is1010jobs available for people. We have been telling people--we have1011been telling kids for to the past 10 or 15 years, go into STEM,1012right, learn to code, like, that is going to be your meal1013ticket. You can have a job for 30 years being a coder, an1014engineer. And now we have AI that can do their job.1015 I talk to a lot of kids who got their IT degrees in1016cybersecurity or different types of technology, and now they1017are having trouble finding a job in this market. So, I would1018love to hear maybe 30 seconds each witness, your thoughts on1019what Congress can do about it, whether we can do anything about1020it at all, whether it is a fixable problem right now. Ms.1021Fabrizio?1022 Ms. Fabrizio. Thank you for the question. I think the1023important thing to continue to focus on is investment in STEM1024education for AI and investment in reskilling and upskilling1025the existing workforce. Apprenticeships will help there.1026 The White House AI and education pledge is looking at these1027areas, and CTA was happy to sign on to that. I will also say1028that, while the workforce will shift, workers will be given new1029tools if they use AI properly, and they will be able to take on1030more capacity, be more efficient, and work smarter.1031 Mr. Subramanyam. Mr. Hammond?1032 Mr. Hammond. I am relatively optimistic about the jobs1033picture over the short run, in particular because I think the1034people who are going to be most displaced are often white1035collar workers who are more adaptable.1036 However, just look at wages for electricians; they have1037spiked dramatically. We have a shortfall in Heating,1038ventilation, and air conditioning (HVAC) and cooling,1039construction, all these things are going--as inputs for data1040centers that could be a major source of job growth. And, more1041generally, I think we need to deregulate aspects of the labor1042markets to make transitions easier.1043 Mr. Subramanyam. What do you mean by deregulate aspects of1044the labor market?1045 Mr. Hammond. Things like occupational licensing, what kind1046of accreditations you need.1047 Mr. Subramanyam. Okay. And Ms. Turner Lee--Dr. Turner Lee?1048 Dr. Turner Lee. Yes. I would say this, on the education1049side, and I do agree with you, Congressman, that we spent so1050much time in STEM and computer science, and since those efforts1051were actually made, we have evolved and sort of retracted on1052those investments.1053 I think we still need to use AI to augment education. We1054often put AI in the classroom instead of talking about the1055education of AI for students, which I think is somewhat of a1056challenge for many teachers and educators.1057 I think it is important to ensure that there is equitable1058distribution of resources that actually train students on AI1059literacy so that they are actually prepared to do more of a K1060through 20 shift as opposed to teaching it in early education1061and then teaching it in college and different levels would1062provide more consistency, more opportunities.1063 And I think that companies need to do a better job of1064qualifying what jobs are going to be lost. I think it is still1065an unknown number of where jobs--companies are going to be1066affected by AI based on the decision of what departments they1067choose to absorb it.1068 And then I also want to just respond to the data center1069side. I think we need to be careful in thinking that data1070centers will generate post-construction jobs and really focus1071on how many jobs will actually be created as a result of the1072data center ecosystem versus what are going to be the job1073creation numbers going into its construction.1074 Mr. Subramanyam. Do you think AI will create more jobs than1075it is displacing, yes or no? Ms. Fabrizio?1076 Ms. Fabrizio. Yes. I think whole new industries will be1077developed because of AI and a tremendous amount of new jobs in1078the workforce will be created.1079 Mr. Subramanyam. Mr. Hammond?1080 Mr. Hammond. When we have fully AI automated software1081engineers, it is less that we lose that job category and more1082that we all become software engineers. And I think that will be1083a general pattern where we are all empowered as individuals to1084take on these new roles.1085 Mr. Subramanyam. Dr. Turner Lee?1086 Dr. Turner Lee. Well, I think if we believe that AGI is1087coming quite quickly that they will take on more of the jobs of1088people because of the superintelligence.1089 I honestly think, Congressman, that AI will change the1090nature of jobs, and that is a conversation that we need to be1091having as opposed to job loss.1092 Mr. Subramanyam. I have more questions, but my time is up.1093We might have a second round.1094 But I will yield back. Thank you.1095 Ms. Mace. All right. I am now going to recognize Mr.1096Burlison for 5 minutes.1097 Mr. Burlison. Thank you, Madam Chair. Thank you for1098conducting this hearing. This is one of my favorite topics.1099 I am going to begin by quoting Irving John Good, who was a1100British cryptologist or mathematician, famous for working with1101Alan Turing. He is quoted as saying, ``Let an ultraintelligent1102machine be defined as a machine that can far surpass all the1103intellectual activities of any man, however clever. Since the1104design of machines is one of these intellectual activities, an1105ultraintelligent machine could design even better machines,1106there would then be unquestionably an intelligence explosion,1107and the intelligence of man would be left far behind. Thus, the1108first ultraintelligent machine is the last invention that man1109may ever need to make.''1110 Mr. Hammond, can you elaborate on some of the success1111recently of some of the AI self-improvement that is occurring,1112where it is constantly improving itself?1113 Mr. Hammond. Yes, absolutely. I think it is coming in in1114gradations. So, we already have AIs that are good at coding,1115and a lot of the job of an AI engineer is coding. And so, there1116is a joke now at these AI labs, they are no longer coding; they1117are just kicking the AI to fix the bugs.1118 Beyond that, we also have AIs that are now writing their1119own algorithms. So, earlier this year, Google released1120AlphaEvolve, which was an evolutionary AI algorithm that1121discovered new bounds on a mathematical theorem that had not1122been beaten in 47 years.1123 Mr. Burlison. And, with that, Madam Chair, I have got an1124article about AlphaEvolve, which is a power coding agent for1125designing advanced algorithms, if I could submit that for the1126record.1127 Ms. Mace. So ordered.1128 Mr. Burlison. Thank you.1129 When--Ms.--Dr. Lee, when it comes to jobs, I am of the1130opinion--and I think you kind of touched on it--that things1131will change. It does not mean there will not be any--there will1132no longer be the need for us to have work to do. I think of it1133often this way.1134 My great-great grandparents, your great-great grandparents,1135probably everyone in this room's great-great grandparents were1136all farmers because that is what it took. It took everyone1137working the fields in order to produce enough food to feed1138people. Today, very few people farm. It is because of1139machinery.1140 And no one would go back in time--would agree that we1141should go back in time and say, ``Do not let them have the1142tractor or the harvester,'' right? If anything, it has taken1143the power or the productivity of one person and magnified it1144manyfold. And that is the way I think we should think about AI.1145 In fact, AI will be a magnifier for productivity for any1146individual no matter what they do and, thus, you need the1147individual. You need that individual that is the core. Would1148you agree with that, Dr. Lee?1149 Dr. Turner Lee. Oh, I definitely agree. I published a book1150last year called ``Digitally Invisible,'' where I went to talk1151to farmers in places in Southern Maryland, across the country,1152and guess what? They said, ``A tractor is only as good as the1153broadband that it has to actually be more productive in the1154work that they do.''1155 But, most importantly, with the compute power of AI, it1156will only be as good as the facilities that we offer them to be1157able to be connected to these new resources, and it will just1158change and transform how they do their work in ways where they1159do not have to go out there and measure how much rain came.1160They will actually know from the comfort of their office that1161is sitting on their land.1162 So, I do agree that we need to have conversations about job1163loss, as well as the transformational capacities in which jobs1164will change the way in which we work. I do not agree that we1165will have robots bossing us around yet.1166 Mr. Burlison. I do not either. I do not either. Ms.1167Fabrizio, the European Union has--they reportedly--their1168regulations have kind of created a chilling effect. Can you1169elaborate on that?1170 Ms. Fabrizio. Yes. It has been hard for companies to1171innovate, and that is why you see fewer unicorns and fewer tech1172innovations out of Europe. I think the United States is doing1173it right, and we have many companies here.1174 We have the most robust startup ecosystem in the world1175here, and we see that firsthand at CES, 1,400 startups, and1176many of them creating new AI innovations and launching them at1177CES. That is why it is really important to make sure that we1178have a framework that supports them.1179 Mr. Burlison. And I think we are at a place now where we1180just realize, if we are going to stay on top competitively, we1181have to be the location for these data centers. We have to have1182these AI, you know, housed in the United States. In order to do1183that, we need electricity.1184 And, Mr. Hammond, would you agree that, right now, we are1185in an electricity crunch? We do not have what is needed for1186that demand, and we have got to change that?1187 Mr. Hammond. Absolutely. Not just the data centers, but all1188forms of things are being electrified from vehicles on down.1189 Mr. Burlison. Thank you. My time has expired.1190 Ms. Mace. Okay. I will now recognize Mr. Crane for his 51191minutes.1192 Mr. Crane. Thank you, Ms. Chairwoman. Thank you guys for1193coming today.1194 I want to start with Ms. Fabrizio. I meet with1195organizations tied to blue collar jobs all the time, and they1196are constantly in my office begging for more help and people1197that need training in specialized fields like plumbing,1198electrical work, and carpentry. Meanwhile, CEOs are warning1199that AI will take 30 percent of the workforce by 2045.1200 My questions are, how do we ensure we are sending more1201people to go to trade schools for these blue-collar jobs and1202not just allowing them to get laid off by AI?1203 Ms. Fabrizio. We need to continue to invest in reskilling1204and upskilling, and we need to make sure that we have the1205resources available. At CES, we are doing AI trainings for the1206first time at CES 2026 for people who are in the industry and1207want to learn more. This is an important solution to this1208challenge.1209 Mr. Crane. Do you agree with those assessments that AI will1210take 30 percent of the workforce by 2045?1211 Ms. Fabrizio. AI is going to change the workforce, but it1212is going to make it better. It is going to create new jobs, and1213it is going to augment the existing jobs. It is going to make1214people smarter. It is going to make them more efficient and1215give them tools that they never had before to do their work.1216 Mr. Crane. Thank you. I would just like to say--make a1217statement for the young generation out there watching, what is1218going on with AI and who may be in college or may be in high1219school.1220 You know, I think one of the fields that is the least1221susceptible to AI taking over and eliminating your career1222opportunities are the trades.1223 What advancements right now are we seeing with technology1224that are creating new job opportunities for Americans?1225 Ms. Fabrizio. Well, we are seeing it in many areas. In1226healthcare, you know, we are seeing individuals that are1227learning how to create autonomous healthcare monitoring systems1228for first-line intake that will help nurses and doctors get1229better information so that they can be physically with patients1230that they need to be with while an autonomous system is1231collecting information on patients. And that would be hugely1232helpful given the healthcare crisis we have.1233 Mr. Crane. Thank you. I want to shift to Mr. Hammond. Last1234week, we lost a true American patriot, Charlie Kirk. Following1235his assassination, there were reports of Chinese and Russian1236bots encouraging violence through spreading misinformation,1237attempting to create division.1238 Did you see any of those reports, sir?1239 Mr. Hammond. I did not.1240 Mr. Crane. Okay. How should the American people be wary of1241the increase of inflammatory speech following the assassination1242of Charlie and other mass violent events from Chinese and1243Russian bots?1244 Mr. Hammond. It is a really big open challenge. These1245social media platforms have their work cut out for them. We do1246not yet have a reliable means of identifying what is a bot,1247what is not, especially as these systems become more and more1248human-like in the way they speak. And so, I think it is1249something we need to put most of our resources into.1250 Mr. Crane. Okay. What do you think Federal agencies should1251be doing to prevent the spreading of misinformation by these1252bots to sow discourse in our communities and our country?1253 Mr. Hammond. I mean, at a minimum, we should stop selling1254China and Russia the technology they use to run those bots. You1255know, these H20 chips, which just got approved or liable to be1256approved for export to China, if they all go through, it is1257going to roughly double their data center capacity for running1258advanced AI models.1259 We know from the past that they have used these chips to1260power their surveillance drones, to power their gate1261recognition technology. So, we have given them the ammunition1262that now they are using on us.1263 Mr. Crane. Next question for you, sir. When I was growing1264up in school, it was often considered cheating to use a1265calculator on a test, right? Now we have CEOs of Fortune 5001266companies basically telling their employees that they need to1267be using AI a few times a day, or they will be falling behind1268or become obsolete.1269 My question is, how do we balance the expanded use of AI1270and not demonize the use of AI while preparing our students for1271the future?1272 Mr. Hammond. I think education is a good example of how AI1273is going to force a massive rethink and reckoning in how we do1274a lot of things, including how we design curriculum for K-12.1275And, you know, there is going to be resistance.1276 But there are already new models that are emerging. There1277is Alpha School in Austin, which is trialing running AI-1278assisted tutoring in the mornings and project-based learning in1279the afternoons and seeing tremendous results.1280 And so, I think we just need much more innovation in how we1281do education.1282 Mr. Crane. What advice for this Committee and for Congress1283do you have in regards to any regulations that you think are1284responsible regarding AI in the future?1285 Mr. Hammond. My three big bullet points are, one, we need1286to monitor the frontier; so we need to know what is coming and1287be able to prepare and adapt because it is going to be a very1288fast-moving period of human history. So, we do not want the1289government to be the last one to know.1290 Number two is going to be investing in research and1291development, especially around issues like control and1292interpretability. How do we interpret how these models work?1293How do we understand their behavior? How do we control their1294behavior? Still, the companies are underinvesting in that.1295 And, third, we need to protect our comparative advantage,1296which is AI hardware. So, as I mentioned earlier, our one big1297advantage is chips and hardware. China is trying to catch up,1298but they are cutoff right now. And, if we open up those chips1299to China, they are going to jump ahead.1300 Mr. Crane. Thank you. I yield back.1301 Ms. Mace. All right. Thank you. We are going to do a second1302round, if that is okay with the witnesses.1303 I request unanimous consent that the Subcommittee have a1304second round of questioning of the witnesses.1305 So, without objection, it is so ordered.1306 I want to pick up where Mr. Crane left off about the chips.1307You are referencing the H20 chip, right, Nvidia? Basically, I1308know China just said that they were encouraging folks not to1309buy from Nvidia, but it was a different chip, right? Do you1310know anything about that?1311 Mr. Hammond. Yes. China is trying to indigenize their own1312chips made by Huawei. DeepSeek was reported to have had some1313botched efforts using the Huawei chips, and so the Chinese1314companies are all hungry for American chips.1315 Ms. Mace. And how do we--I mean, how do we prevent China1316from getting the chips?1317 Mr. Hammond. First order is we should not approve the sale1318of the chips. And, to the extent that we do, try to minimize1319the damage. Senator Jim Banks has a bill called the GAIN AI1320Act, which would give U.S. companies a first right of refusal1321to buy the chips that are destined for China. I think that is1322the least we could do.1323 Ms. Mace. Is there a House version of that bill?1324 Mr. Hammond. I do not believe so, not yet.1325 Ms. Mace. Okay.1326 Mr. Hammond. You know, even with these controls in place,1327China has very sophisticated smuggling operations. So, not a1328month after the Blackwell series of chips were announced, there1329was an FT report that China had already smuggled in a billion1330dollars' worth of them.1331 So, we need to do much more to crack down on----1332 Ms. Mace. How did they do that? How were they able to get1333away with that?1334 Mr. Hammond. Through third-party intermediaries in1335Malaysia, Singapore, Taiwan. They just--they buy them through1336these third parties that--on the list. In some cases, they are1337even incorporated in the United States, and it is just a matter1338of getting them over the border.1339 Ms. Mace. And are we cracking down on the intermediaries1340here in the United States now? Have you heard anything about1341that?1342 Mr. Hammond. There have been some DOJ-style investigations1343that there have been actions taken in Malaysia and Singapore.1344We have enlisted the governments in those countries in some1345cases.1346 The challenge is really one of scale. So, once these chips1347are at the door, how do we know where they are ending up? There1348is a bill called the CHIP Security Act. There is a House1349version of that that would require these chips to have basic1350location verification.1351 Ms. Mace. Who is doing that bill?1352 Mr. Hammond. It is before the House Foreign Affairs1353Committee (HFAC) right now.1354 Ms. Mace. Okay. And then--so, Ms. Fabrizio, you said1355earlier in one of the questioning about China, using AI1356differently. Can you give some examples?1357 Ms. Fabrizio. Well, China is using AI for surveillance,1358they are using it for military purposes, they are--you know,1359they also do not have privacy in China. So, they use the data1360differently than we do here.1361 That is why I think it is really important that we continue1362to move forward, we look for a national AI framework that1363addresses some of these big issues. It is risk-based, It is1364tech neutral, and it allows innovation to continue to flourish1365here so we can continue to beat China.1366 Ms. Mace. Is China our greatest threat?1367 Ms. Fabrizio. Yes.1368 Ms. Mace. Do you agree, Mr. Hammond, China being the1369greatest threat?1370 Mr. Hammond. Certainly on the chip----1371 Ms. Mace. Dr. Turner Lee, would you agree?1372 Dr. Turner Lee. Yes.1373 Ms. Mace. What are the consequences--and this is for all1374the panelists--if the United States fails to outpace China in1375the race for domination in AI? Ms. Fabrizio?1376 Ms. Fabrizio. Well, it is important that we move forward in1377the best way that we can and that we focus on innovation. We1378have the best startup ecosystem here. We have the best tech1379companies here, but we need to not get in their way.1380 So, that means a pause on state legislation. It means a1381Federal framework where companies can innovate and know the1382guardrails and be able to move forward.1383 Ms. Mace. Mr. Hammond?1384 Mr. Hammond. I worry that we need to buy time with China1385because over the last 40 years--this is part of the1386industrialization story--we have shifted all our industries1387into services, entertainment, law, finance, all things that are1388about to be deflated by AI whereas--so there is a world where1389we build the AGI, we build the general intelligence, but China1390is the one that puts in factories and has the growth benefits.1391 Ms. Mace. There have been people--because I know one1392personally, someone I am suing--literally, he is using ChatGPT1393as his attorney. I just cannot think of any--I mean, okay, yes.1394But--we are going to beat the LLM in court. Dr. Turner Lee?1395 Dr. Turner Lee. I would agree with my colleagues in terms1396of China being a threat to our dominance in the AI space. I1397also would put out there, though, that we have to find1398alternative markets if we are actually going to grow the1399economy and scale of U.S. companies.1400 So, I am thinking about my experiences with 5G, where we1401actually opened up other markets, we missed the opportunity to1402work with the global majority, African Union. So, just1403thinking--rethinking our industrial policy as we think about1404China as a threat as to if we want people to have American1405products embedded in their technology, where are we selling it1406to, and making sure we are agile.1407 I do also want to respond that I do think states have to1408play a role in experimentation. I think the Federal Government1409would be too premature to come up with a national policy that1410limits states' rights because what we are seeing in terms of1411experimentation of states is that they are looking at more1412consumer protection.1413 AGs are trying to figure out ways to keep our grandmothers1414safe from AI. They are not necessarily trying to compete1415against China. They are just competing against the various1416misnomers and lack of information that people----1417 Ms. Mace. We already have that. I got a Nigerian scam the1418other day in an email that, you know, I took a screen shot and1419forwarded to my family, and I am, like, ``Don't click on1420this.''1421 Dr. Turner Lee. We do the same thing because we are not1422coming to Congress; we are going to our state AGs. So, we can1423eval what they are able to do to make sure people are safe.1424 Ms. Mace. Yes. Okay.1425 I will now recognize Mr. Subramanyam for 5 minutes.1426 Mr. Subramanyam. Thank you, Madam Chair.1427 I just want to finish the job displacement conversation,1428and, Mr. Hammond, you said something interesting about the1429United States has shifted its economy toward industries that1430are susceptible to AI. You mentioned law. I am a lawyer too,1431and so I am sensitive to that, but can you expound a little bit1432on that as well as what should we be doing? I know Mr. Crane1433made some good points as well about, you know, different1434industries that might be more important moving forward, but I1435would love to hear more about that.1436 Mr. Hammond. It is a story of relative scarcity. So, if1437intelligence becomes abundant, if service labor, cognitive1438labor becomes abundant, then what remains scarce? And it will1439be the heavy industry, the factories, the actual--not just the1440factories, but also the know-how, the tacit knowledge that is1441embedded within the workforces that China has that we do not.1442And so, you know, one of the reasons that we need to buy time1443is in part to try to rebuild some of those new sectors in part1444to ensure that we do not just give away our economy.1445 Mr. Subramanyam. And what would buy time? What can we do to1446buy time?1447 Mr. Hammond. Well, number one, denying China access to the1448most advanced AI chip.1449 Mr. Subramanyam. Oh, I see. Okay. Got it. You mean like1450export controls, and that sort of thing. I would ask, I guess,1451the other witnesses as well what your thoughts are on this?1452Like, what jobs are we going to lose in the future as well?1453What jobs are we going to be gained? Ms. Fabrizio mentioned1454this is going to create more jobs than we are losing. Do we1455even know right now what those jobs would be? I know it is--you1456know, there is this idea of, okay, well, AI is going to1457displace this IT department but it is going to replace it with1458AI buddies that, you know, help fix the AI or make it better.1459But, you know, I look at what is actually happening and that is1460not quite the reality right now. Perhaps it will be in the1461future, but right now what I am seeing is, you know, 40 people1462get laid off in an IT department. They are replaced with AI1463and, like, five people, and so I am just curious what you think1464that future would look like, what jobs will be gained.1465 Ms. Fabrizio. Well, every major shift we have seen new1466industries and new jobs created. Think about the internet.1467Things--you know, so many new industries were developed, and we1468see that at CES. You know, in terms of jobs, there are AI data1469scientists, AI ethicists. There are new ways of building and1470manufacturing technology that did not exist before, and we will1471continue to see more and more of that as new industries use AI1472and as new--whole new industries that we do not even know what1473they are yet emerge. But agriculture is one great example where1474we have seen new roles develop because of AI.1475 Mr. Subramanyam. Ms. Dr. Turner Lee?1476 Dr. Turner Lee. Yes. And I would just say, in sort of1477looping back to the previous question as well, I think we have1478to be distinguishing between the loss of jobs in the blue-1479collar sector and the white-collar sector, right? So, I think1480on the white-collar sector it has been very clear that AI is1481going to improve the efficiencies of lawyers or paralegals who1482do the type of research that AI can do expeditiously. On the1483blue-collar side, it becomes a little bit more tricky, because1484when we are talking about trade jobs that are going to be lost.1485I am not sure yet if a robotic plumber can come to my house and1486fix my toilet, but we still have to see that we are going to be1487running these parallel workforce opportunities, and it is1488important for that plumber to have the skills necessary to be1489able to innovate and to grow into the new economy where maybe1490they are not actually doing the physical labor, but they are1491managing schedule, or they are managing invoices, or they are1492trying to do trade service calls. I want to put that out there.1493 I think the question that we all should have is we are1494going to lose collective bargaining jobs when AI comes in,1495because it is going to replace front line workers who1496essentially have had that job security as well as a union to be1497able to do the work that they do. So, I think we have to just1498look at the scan and do a better scan of what that means when1499we say job loss--Congressman, you are right on point--and sort1500of divide that out based on not only productivity but where we1501are going to see the most vulnerability.1502 Mr. Subramanyam. And I would ask all the witnesses what1503should Congress do, if anything, about job displacement? Ms.1504Fabrizio?1505 Ms. Fabrizio. I think investing in workforce development1506and training and apprenticeships. I mean, you are right, you1507might not have a robotic plumber, but you might have a plumber1508that is able to look at a digital twin of your home and1509identify a better solution faster that they would not have had1510the ability to do before. So, continuing to invest in1511upskilling is extremely important.1512 Mr. Subramanyam. Mr. Hammond.1513 Mr. Hammond. I would say two things. One, that the U.S.1514Government has a poor track record of running employment and1515training and retraining programs. They tend to not work very1516well. And I believe GAO last reported that there are 46 of them1517already. And so, I would look to opportunities that shift1518training on the job as much as possible, and, to the extent1519possible, reducing barriers to enter new jobs so that you do1520not need to get that certificate or do not need to get the1521extra piece of education just to enter the workforce.1522 Dr. Turner Lee. And I would say the same thing. Upskilling,1523obviously, cross-skilling as the nature of jobs actually are1524transformed. I love the example of my colleague with regards to1525digital twins in some of the trade areas. And I also would say1526national AI literacy so that people also understand that AI is1527not just the job. It is the behavior in which you approach the1528task that is before you. That is one of the most transformative1529aspects of this technology compared to any internet1530technologies that we have had. You can actually see AI through1531an appliance and figure out ways in which you are interacting1532with it. As a student, you can see it on your phone. So, really1533thinking about what does AI literacy look like so that people1534understand how that actually fits into the labor force.1535 Mr. Subramanyam. Thank you. I yield back.1536 Ms. Mace. Thank you.1537 Mr. Burlison, would you like to be recognized for 51538minutes?1539 Mr. Burlison. Sure. Thank you. Ms. Fabrizio, at one point1540in your, I think in your, testimony you talked about that,1541without Federal privacy law, businesses and consumers are1542navigating the confusing state by state patchwork. I could not1543agree more. But can you elaborate on what--by us not having1544some kind of standardized laws in place and really kind of open1545the door for one state to squelch this industry.1546 Ms. Fabrizio. Well, yes, and more so it is very harmful to1547businesses, especially small businesses or startups with a1548unique idea to try to figure out how to scale. It is very1549stifling to try to comply with many different state laws and to1550grow your market and to grow your product, and so that can be1551discouraging and that can discourage innovation, and that is1552not what we want to do here. We want to encourage innovation,1553and a Federal privacy framework would help give some1554consistency and some clear guardrails and rules of the road so1555that our innovators can innovate.1556 Mr. Burlison. Right. Do you think we need in general, just1557more broadly speaking, privacy laws related to our own personal1558data?1559 Ms. Fabrizio. Well, in terms of looking at a framework for1560privacy or for AI, it should be risk based, and we should1561approach it that way, but I do think, versus a patchwork1562approach that, again, companies just cannot adhere to it. They1563cannot. Small companies cannot afford big privacy law firms to1564support them. That gives the big companies a competitive1565advantage and boxes out small companies.1566 Mr. Burlison. So, with that being said, what would you like1567to see as a framework for setting wide sweeping Federal law?1568 Ms. Fabrizio. So, we would be looking for something that is1569tech-neutral, that is preemptive--it is one framework--and that1570it is risk based, and that it identifies those core categories,1571and it also, you know, removes liability for companies that are1572compliant, and that I think would really unlock and propel1573innovation and move us forward.1574 Mr. Burlison. Interesting. You were speaking kind of1575esoterically or a little bit philosophically there, but1576specifically, can you think of specific rules or regulations1577that you would want to see implemented?1578 Ms. Fabrizio. I would love to get back to you with1579specifics and have a further conversation about it.1580 Mr. Burlison. I respect that. Thank you.1581 Mr. Hammond, a lot of this debate is about what the values1582that we are going to teach this AI, which I find interesting1583that we cannot even agree on what free speech means as1584humanity; how can we trust AI to determine or how can we1585attribute or provide some kind of values to AI?1586 Mr. Hammond. This is an unsolved problem as it stands. So,1587you know, right now, you know, chat bots will adopt whatever1588value you put into it. If you tell it to talk like a pirate, it1589will talk like a pirate. The question is when these systems1590become more autonomous and acting on their own, how do they not1591just follow our values but follow them reliably. I think that1592is an area still needed for further research.1593 Mr. Burlison. Thank you. Because I am touching back on the1594question I asked Ms. Fabrizio, which is that I think that the1595vulnerability for each one of us is that we are entering into1596an era where an AI can devour as much information about you as1597possible and weaponize that against you, and so that is1598something I think is a concern, and the question is how do we1599protect people's rights?1600 For example, every day I get annoying phone calls for1601people from, like, people that want to know if I want to sell1602my rental homes. Every day. I never signed up for anything.1603Somehow they know which homes I have, and they just always want1604to call me. I can only imagine how bad that is going to get1605when--how much worse it is going to get, which is why I think1606we should be considering as--on a Federal level having a1607tighter grip or control on the data about individuals,1608particularly important data, whether it is the electronic1609medical record. We may need to do something like what we did1610with Health Insurance Portability and Accountability Act1611(HIPAA) or extend that beyond just your medical privacy. Dr.1612Lee, you are nodding your head.1613 Dr. Turner Lee. I am nodding my head. We are way overdue in1614the United States for a national privacy standard that would1615actually dictate what goes into the machines, what comes out of1616it in terms of what we consent to, not consent to implicitly as1617well as, you know, the ability of third parties to get ahold of1618that. You get calls about your rental properties. I am offered1619$100,000 every day through some type of voice clone. So, I1620think that, alongside deep fakes, which actually only exploits1621the opportunities you are talking about. We really do need1622Federal privacy legislation to slow this down.1623 Mr. Burlison. Thank you. I yield back.1624 Ms. Mace. All right. I will now recognize Mr. Crane for 51625minutes.1626 Mr. Crane. Thank you.1627 I want to get back on this chip conversation that we were1628having, Dr. Hammond. You were talking about the need to buy1629time. I believe we were talking about the H20 chip. Is that1630correct?1631 Mr. Hammond. Yes, sir.1632 Mr. Crane. Are those manufactured in Taiwan or the United1633States?1634 Mr. Hammond. Primarily Taiwan.1635 Mr. Crane. Do you have fears that--China has stated1636repeatedly that their plans are to take Taiwan, and by doing so1637would have access and control of all of these chip companies1638that the United States has invested in?1639 Mr. Hammond. I am very concerned about a potential invasion1640or blockade against Taiwan. I would say that, if that were to1641occur that Taiwan Semiconductor Manufacturing Company Limited1642(TSMC) would not be long for the world.1643 Mr. Crane. On that note, just to unpack it for the American1644people that might be watching, what percentage of chips that we1645use as American consumers and in AI come from Taiwan?1646 Mr. Hammond. When it comes to the most advanced AI logic1647chips like the products Nvidia produces, over 90 percent.1648 Mr. Crane. 90 percent.1649 Mr. Hammond. With the new TSMC factories going up in1650Arizona and elsewhere, we are going to try to grow our share,1651but right now it is----1652 Mr. Crane. Is it not, like, 60 percent of semiconductors1653that we use come from Taiwan as well?1654 Mr. Hammond. Yes. Across the board.1655 Mr. Crane. What do you think would look like if that1656scenario were to unfold within the next couple years, what do1657you think--effects that would have on the U.S. economy?1658 Mr. Hammond. I think we would have a global depression,1659number one, and then number two, all our products from our car1660door to our toaster have semiconductors in them.1661 Mr. Crane. When I have heard experts talk about, you know,1662standing up these manufacturing capabilities that, you know,1663building these advanced chips, it is not easy. It often takes1664decades. Would you agree with that?1665 Mr. Hammond. It can from scratch, but the fab going up in1666Arizona is actually ahead of schedule and showing very good1667results after only a couple years.1668 Mr. Crane. So, you think, within a couple years, we can be1669producing these most advanced chips right here in the United1670States?1671 Mr. Hammond. If similar style, like if a CHIPS Act 2.0 were1672to come across, we could do this again, yes.1673 Mr. Crane. Ms. Fabrizio, you talked about other eras within1674history, like the invention of the internet, and you compared1675that to AI to kind of make your point that, you know, any time1676there has been serious innovation within industry, it often1677leads to new jobs, other fields that we do not even yet know1678about. Do you think that that is coherent and fair looking at1679what we are facing here with AI? Do you think it will be1680similar, or do you think AI will be a lot more disruptive for1681jobs in the economy than anything we have ever seen?1682 Ms. Fabrizio. I think it will be a good shift that we see.1683There will be disruption, but I think it will be positive. I1684think it will allow us to solve some of the biggest challenges1685that we have. Think in healthcare: we have a healthcare worker1686shortage. We have more sick patients. Think food insecurity,1687think farming, agriculture, energy, mobility, smart cities. I1688see all of these solutions at CES. I see amazing groundbreaking1689technologies, and I would invite you all to come to CES and see1690them, too, and see how the future will be amazing once we1691continue through this shift if we allow the United States to be1692the leader and continue to be the leader.1693 Mr. Crane. Do you agree with that, Mr. Hammond?1694 Mr. Hammond. Yes. I would add a component to this. A vital1695component is cybersecurity, infrastructure security. So, it is1696one thing--we need these data centers in this country first and1697foremost for sovereignty if we are going to be running through1698the economy through these data centers. If they are going to be1699contributing to GDP, they should be within our borders, within1700our jurisdiction, but that also creates a single point of1701failure, right? If the power goes out today, we are still able1702to have a conversation; but, in the future, if the world is1703running on AI, that is a huge critical piece of infrastructure1704that could be taken out.1705 Mr. Crane. I think we did a hearing on this same topic1706probably a couple months back, and one of the AI titans, you1707know, we brought in an article, and he said within the next one1708to five years, like, 50 percent of all entry-level white-collar1709jobs are going to be gone. So, you think those folks are all1710going to be able to find new jobs?1711 Ms. Fabrizio. Well, I do not think those jobs will be gone.1712I think there will be different jobs, and those people1713hopefully will be focusing on upskilling. They will be using1714AI. We have all heard that AI will not replace people, but1715people who use AI will replace people who do not, and so that1716is how I look at it, and I do think that is the shift that we1717will experience.1718 Mr. Crane. Okay. So, you guys do not think this is going to1719be, like, one of those times where we told coal miners, ``Hey,1720when your plant goes under and we get rid of it, we are going1721to teach them how to code?'' Do you guys remember that a couple1722years ago? What do you think those coal miners are thinking1723right now? How many of them do you think learned how to code?1724 Mr. Hammond. Teach coders how to mine.1725 Mr. Crane. I am not saying it is your responsibility. I am1726just saying this is one of the biggest concerns that I have1727when I look at what is going on with AI.1728 Ms. Fabrizio. I agree. I think upskilling and investing in1729the future workforce is extremely important, and it is1730something that we have to do, and we have to bring people1731along, but that involves education. It starts with STEM1732education early in AI and continued investment there and being1733future looking.1734 Mr. Crane. I get it, but a lot of these folks that already1735have these jobs. They already got their education that, in many1736cases, are going to be worthless. Would you agree, Dr. Turner--1737or----1738 Dr. Turner Lee. Oh, of course. I think what you are1739actually pointing out, Congressman, is that we need to do more1740research on this, and we need a little bit more data before we1741jump in here and start asserting what we think the workforce1742will look like.1743 Mr. Crane. Thank you. I yield back.1744 Ms. Mace. Okay. In closing, as we wrap this up, I want to1745thank our panelists once again for being here this afternoon1746and providing your testimony. I would like to yield to the1747Ranking Member for any closing remarks.1748 Mr. Subramanyam. Thank you, Madam Chair.1749 I think some really interesting points were made by Members1750of the Committee today, and I just--I want people to understand1751that we should not downplay the job losses, the job1752displacement that is already happening. This is not1753theoretical. I had a job fair in my district a couple months1754ago, and I was ready to meet many folks who had lost their jobs1755because they were Federal workers, and I have a lot of Federal1756workers in my district that have been laid off, and certainly1757there were those. But there were even more IT students and1758graduates who were there because they had lost the ability to1759get a job because companies were not hiring them anymore. They1760wanted ten years of experience. Well, if you are an IT student1761who just graduated from a 4-year college, you were told your1762whole life to go into STEM, and you got the best STEM education1763from the best STEM education schools, and now there is no job,1764and they are asking me why; what happened?1765 And then I look at what is happening at the companies, and1766they are laying off entire departments and replacing them with1767AI and putting out press releases in some instances, bragging1768about it to their shareholders. Again, I am not here to run1769their business. I think it is reality. If you could replace 1001770people that you spend a million dollars paying with a tool that1771costs $50,000, it is part of your mandate to do that, but I1772think we have to understand that this is an imminent problem,1773that certainly we have invested a lot already in STEM education1774and in job training programs.1775 But I actually kind of agree with you, Mr. Hammond, we have1776a lot of them already. I do not know if that is the only1777solution here. I do not have any answers for you today. That is1778why I was asking these questions, but I do want to work with my1779colleagues on both sides of the aisle to figure out what comes1780next, because I have a lot of students with really good skills.1781They do not need to be upskilled anymore. They have really good1782tech skills. They just cannot find jobs, right? And so, I want1783to see if there is something we can do about that. I yield1784back.1785 Ms. Mace. Thank you.1786 And, with that, without objection, all Members will have1787five legislative days within which to submit materials and to1788submit additional written questions for the witnesses which1789will be forwarded to the witnesses for their response.1790 And, if there is no further business, without objection,1791the Subcommittee stands adjourned.1792 [Whereupon, at 3:20 p.m., the Subcommittee was adjourned.]17931794 [all]Witnesses
3 witnesses appeared, with 9 papers on file.
- Witness Biography — HHRG-119-GO12-Bio-TurnerLeeN-20250917.pdf
- Witness Truth in Testimony — HHRG-119-GO12-TTF-TurnerLeeN-20250917.pdf
- Witness Statement — HHRG-119-GO12-Wstate-TurnerLeeN-20250917.pdf
Documents
The committee filed 5 documents for the meeting.