"Building an AI-Ready America: Higher Education in the Age of AI"
Legend
Transcript markers
- text Sentence being played
- text Search result
- text Intervention cited by the link you opened
Click a sentence to play the video from that moment. Select text to quote it with the speaker, timestamp and link.
More options
House Education and Workforce Subcommittee on Higher Education and Workforce Development U.S. House of Representatives
Follow its sittingsAlma Adams
We need clear guardrails to ensure AI strengthens education rather than weakens it. Unfortunately, the current administration's policy is moving us in the wrong direction. At a moment when stronger oversight is needed, we are seeing hesitation, fragmentation, and the weakening of the very institutions responsible for protecting students and educational professionals. As private companies rapidly collect data and expand AI systems into classrooms and campuses, students and institutions are increasingly at risk. And what makes this moment especially concerning is that the Trump administration is weakening the very institutions responsible for protecting students while simultaneously pushing to rapidly expand the use of AI.
This administration says it wants to unleash AI. Yet it has supported efforts to dismantle and weaken the Department of Education, reduce capacity within the Office of Civil Rights, eliminate the office responsible for educational technology guidance and support, and has scaled back workforce preparation, including the Office of Career, Technical, and Adult Education. That is deeply counterproductive. These gaps matter because AI systems are not neutral. They can reinforce bias, misuse student data, and create discriminatory outcomes that directly impact access to education, financial aid, disability accommodations, and academic opportunity.
And without strong federal oversight, students, faculty, and educational workers are left vulnerable with little recourse when their rights are violated by these systems. And that is why it is our job to step in. We need strong guardrails and infrastructure that match the scale of this technological shift. And that means restoring and strengthening federal educational technology leadership. It means rebuilding OCR's capacity so that complaints involving AI discrimination and data misuse can actually be investigated.
It means establishing clear accountability standards around transparency, data protection, auditing, and oversight for AI systems used in higher education. But most importantly, higher education institutions need coordinated federal guidance instead of being left to navigate this alone, campus by campus. AI is already here. Ignoring it will not stop it. But failing to govern it responsibly will deepen inequality and weaken trust in education and leave students unprotected during one of the most significant technological transitions of our time.
We have an obligation to ensure that AI is used ethically and responsibly to serve students, educators, and the public interest. And with that, Mr. Chairman, I yield back.
Burgess Owens
Pursuant to Committee Rule 8C, all members who wish to insert written statements into the record may do so by submitting them to the committee clerk electronically in Microsoft Word format by 5:00 PM, 14 days after this hearing. Without objection, the hearing record will remain open for 14 days to allow for such comments and other materials noted during the hearing to be submitted for the official hearing record. I would now like to turn to introductions of our four distinguished witnesses. Our first witness is Jonathan Fozard, Chief Information Officer at Florida State University in Tallahassee, Florida. Our second witness is Dr. Dave Duke, Chief Product Officer for Higher Education at McGraw Hill in Chicago, Illinois.
Thank you.
The third witness is Dr. Bridget Burns, CEO at University Innovation Alliance in Tempe, Arizona, sorry. And our fourth witness is Michael Horn, an author and adjunct professor at Harvard Graduate School of Education in Cambridge, Massachusetts. We thank the witnesses for being here today and look forward to your testimony. Pursuant to committee rules, I will ask each of you to limit your oral presentation to a three-minute summation in your written statement. The clock will count down for three minutes as committee members have many questions for you.
However, pursuant to Committee Rule 8D and committee practice, we will not cut off your testimony until you reach the five-minute mark. I would also like to remind the witnesses to be aware of your responsibility to provide accurate information to the subcommittee. I would like to first recognize Mr. Fozard for your testimony.
Jonathan Fozard
Thank you, Chairman. Chairman Owens, Ranking Member Adams, and members of the subcommittee, thank you for the opportunity to testify today on an important topic of AI-ready America. I am honored to appear before you and grateful for the opportunity to share the higher education perspective on one of the most consequential issues facing our country. My name is Jonathan Fozard and I am the Chief Information Officer at Florida State University. Artificial intelligence is not simply another emerging technology.
It is quickly becoming a defining capability that will shape the economic competitiveness, national security, scientific discovery, education, healthcare, public service, and the future of our work. The United States right now is a global leader in artificial intelligence, and we must remain at the forefront. This is a pivotal moment. Institutions, organizations, companies, and nations have a choice. They can embrace AI responsibly, or they can choose not to use it at all.
That decision will increasingly differentiate who is leading and who is falling behind at an exponential rate. That is why education must be central to America's AI strategy. America's AI leadership will not be secured by technology companies alone. It will require national education and research pipeline that begins in K-12 and continues through higher education, expands into the workforce training and lifelong learning, and is strengthened by meaningful partnerships among university, government, and our industries. At Florida State University, we see this responsibility very clearly.
Higher education must prepare students not only to use AI, but to understand it, to question it, to improve it, secure it, and apply it in ways that serve people and strengthen our nation. I have many waves of change and have seen many waves of change as a technology practitioner for more than 26 years. As a father of two children who will soon enter the workforce, this topic is also very personal to me. I want my children and students across this country to inherit a nation that continues to lead, continues to innovate, protect its values, and create opportunities for generations to come. I believe that five priorities should guide our nation's approach as we move forward.
First, I believe that we must treat AI literacy as a national workforce priority. Second, we must support secure and responsible access to AI tools so that students are prepared for the future. Third, we must invest in hands-on learning and workforce development programs. Fourth, we must strengthen the K-12 to higher education pipeline.
We are doing this because AI readiness is not optional. It is central to our students' future, our workforce, our research enterprise, our national competitiveness, and our responsibility to the country. The United States has the talent, the institution, the innovation ecosystems, and the values needed to lead in a world of artificial intelligence. But leadership is not guaranteed. It must be built, protected, renewed, sustained, and invested in education, research, infrastructure, security, and most importantly, people.
This is the moment for urgency, but also for optimism. Higher education stands ready to help build the next generation of AI leadership and innovation. Working together, we can ensure that the United States remains at the forefront not only for our own future, but for the future of our allies, our partners, and the democratic values we all share. Thank you for your time, for your service to our country, and for your commitment to building this AI-ready America. I look forward to your questions later.
Dave Duke
Thank you. I want to talk about something more uncomfortable than AI adoption rates, a gap that is opening in American higher education right now that our institutions are not yet equipped to close. Two things are happening simultaneously. Employers across every major sector of the American economy now expect AI proficiency from college graduates. It's showing up in entry-level job descriptions.
And hiring managers, many of which are not yet adept at using AI themselves, are expecting graduates to work with AI tools productively and responsibly within a specific professional domain. And this is also true: most institutions of higher education are not preparing their graduates to meet that expectation. The response to AI on campus across the country has been fragmented and inconsistent. The gap between what higher education is producing and what the economy now requires is real, and it is growing, and it is a structural challenge. And I call this the proficiency dilemma.
And it has two sides that are in tension. On one side, students are using AI constantly. They've developed a practical fluency through daily use. But fluency through unsupervised use is not the same as cultivated professional competency. They've not developed the critical thinking or domain expertise that AI will demand of them as professionals, and they have learned to produce outputs without developing the capacity to evaluate those.
On the other side, academic institutions that have responded to AI primarily through restriction and detection are producing a different kind of graduate altogether, one who has been taught that AI is something to be managed and avoided rather than understood and used with skill and judgment. And so the right answer is neither unrestricted use nor aggressive restriction. It is intentional, structured, pedagogically sophisticated integration, teaching students not only how to use these tools, but when, why, and with what critical judgment. Very few institutions are doing this well. And I want you to know that.
Faculty are making individual decisions in a policy vacuum, often without guidance, without professional development, and without the time to think through what responsible AI integration requires. And let me explain what is at stake if we get this wrong. We will produce a generation of graduates who are simultaneously over-reliant on AI and under-prepared to work with it professionally. We will widen, not narrow, the gap between students who receive structured AI education and those who do not. And we will perpetuate a disadvantaged workforce that has a difficult time competing in the global economy.
Higher education has always been America's mechanism for translating individual potential into national strength. And if we allow it to fall behind the economy it is meant to serve, we are not just failing students, we're failing ourselves.
And I'll leave you with three recommendations. First, treat AI proficiency as a core educational competency, a nationwide commitment built into accreditation standards and general education frameworks to clarify the expectation of an AI-ready America. Second, initiate the development of institutional AI frameworks that create actionable guidance that faculty and administrators need but currently lack so that individual educators are not left to navigate this transformation on their own. And third, examine the variation in AI readiness among all institutions. Explicitly well-organized and better-positioned universities have advantages in AI adoption that smaller regional and community colleges do not, and the benefits of AI in higher education should be broadly shared.
And so the question before American higher education is not whether AI will transform the experience of learning and working. It already has. The question is whether our institutions will lead that transformation. The students sitting in classrooms today will spend the most consequential years of their careers in an AI-shaped economy. They deserve an education that prepares them for that reality.
McGraw Hill is committed to that work, and we're grateful for the subcommittee's attention to it. Thank you, and I welcome your questions.
Bridget Burns
Chairman Owens, Ranking Member Adams, and members of the subcommittee, thank you for the opportunity to testify today. My name is Bridget Burns, and I serve as the CEO of the University Innovation Alliance, which is a national consortium of 19 large public research universities across the country serving more than half a million students. Over a decade ago, the UIA was formed by our presidents and chancellors who were trying to respond to a national challenge. We needed to increase the number of college graduates to strengthen the economic competitiveness of our country, and over the past 12 years, our member institutions have stepped up and produced more than 180,000 additional graduates above their stretch capacity in that time while simultaneously increasing their graduates from low-income backgrounds by 51%. We did this by doing something higher education rarely does, which is we innovated together and we shared openly what we learned in real time.
Our campuses have leveraged AI, machine learning, chatbots, and predictive analytics to improve student outcomes for more than a decade. That experience directly shapes how we view artificial intelligence. It's tempting to treat AI as a technology challenge, but it is far more all-encompassing than that. It is rapidly reshaping teaching and learning, workforce preparation, research, advising operations, and student support. In less than two years, institutions have gone from asking whether a student would use these tools to manage a reality where their use is every day.
One of the reasons navigating AI transformation is so challenging is that it does not neatly fit into any traditional organizational structure. It cuts across every aspect of the institution. Institutions across the country are trying to answer the exact same questions responsibly around workforce preparation and privacy and governance and learning, but most institutions are not actually resisting AI. They are trying to navigate it, but most of them are navigating it alone. That is the problem.
My concern is that we are approaching AI primarily as an institutional challenge when it's increasingly becoming a national competitiveness challenge.
If America wants to lead in an AI-enabled future, we need to be organizing around national priorities and leveraging the full capacity of our higher education institutions to help achieve them. That requires coordinating infrastructure that does not exist. Right now, thousands of institutions are independently evaluating vendors and developing governance frameworks, piloting tools and creating policies. And too often, the lessons learned are remaining in one place and not going to another institution. We're duplicating effort at a moment when we should be accelerating learning across the country.
We also need to not leave under-resourced institutions behind. Whether a student benefits from responsible AI adoption should not depend on where that student happens to enroll. Students attending rural institutions, regional universities, community colleges, HBCUs, all of them deserve access to the same opportunities at the wealthiest institutions in the country. America's economic competitiveness depends on developing talent everywhere. Our campuses are already demonstrating what responsible AI adoption looks like.
At Arizona State University, for example, they developed something called the Triangulator, which streamlines transfer credit evaluation, a process at any other institution that traditionally requires staff manually weeks to compare catalog course by course. The tool uses AI to identify likely equivalencies and present recommendations for staff review, with a human expert always making final decisions. AI open-sourced this solution and already many institutions, including seven of my own, are working to scale it. Purdue University was the first institution in the country to integrate AI literacy into their graduation requirements. At UC Riverside, the writing faculty have designed a Socratic tutoring tool that coaches students through the work via questioning rather than rewriting.
It preserves authentic learning while providing the personalized learning and support they need. And at University of Utah, they're piloting AI-enabled tools that help students connect academic pathways to career opportunities. Importantly, these institutions are not treating AI innovation as a proprietary competitive advantage. They are openly sharing what they learn so other institutions can adapt and benefit from the work. What concerns me is that we are doing so while campuses are operating under significant strain.
Leadership turnover, staffing shortages, financial pressure, and uncertainty around research infrastructure are making it harder for institutions to build the capacity required to navigate this moment successfully. Federal policy has spent decades rewarding institutions for developing innovation. The AI era requires us to also reward institutions for sharing them. One of the most effective ways to do that would be to prioritize multi-institutional partnerships in federally funded AI initiatives. I'm looking forward, thank you so much for your questions and I look forward to your time.
Michael Horn
Thank you. Lastly, I'd like to recognize Mr. Horn for your testimony.
Chairman Owens, Ranking Member Adams, and members of the subcommittee, thank you for inviting me today. AI alone will not improve higher education. My central message today is simple. What matters is whether colleges redesign their resources, processes, and priorities around it. The analogy is electricity.
Early factories adopted electric motors but saw little productivity gain because they kept the same organizational designs and processes in place. Only later, when they redesigned themselves around electricity, did productivity surge. Higher education is at a similar moment. Adoption has, in a sense, been quite rapid. Today, faculty and students are already using AI at scale.
Yet most colleges, with a few exceptions named earlier, have seen few learning leaps, few dramatic improvements in student success, and few efficiency gains. Most campuses remain somewhere between a wait-and-see approach and a patchwork of siloed experiments. The outcome is frankly predictable. In every sector, when organizations treat the adoption of technology as a problem of technology, progress is slow and uneven.
Indeed, organizations that frame the challenge as one of technology often perceive it more as a threat than an opportunity. Hence how AI is exacerbating cognitive automation on the part of students and faculty overwhelmingly report concerns around plagiarism and cheating. The challenge is actually the operating model itself and the opportunity is redesign. We can in fact use AI to increase academic rigor and reinvent assessment. If AI can complete an assignment, perhaps the assignment in itself is in need of change.
Oral defenses, presentations, demonstrations, and rigorous feedback can help ensure that students are still doing the hard work of learning. We can use AI to boost student success by connecting data around academics, finances, and well-being to offer more meaningful support. We can also streamline processes and reduce administrative overhead. And we can better connect learning to work. As AI rapidly changes jobs and the half-life of skills shrinks, students need opportunities to do real work in real-world settings and use AI in the course of that work.
Employers are not asking whether employees should use AI, they are expecting it. That means that schools must offer embedded projects and simulations in courses, externships, paid internships, co-ops, or apprenticeships. All the better when these experiences both pay, count for credit, and connect students to working professionals to help them build social capital. Many universities are stepping up on this front, from Northeastern University with its co-ops to Minerva University, where I'm a trustee, with its combination of online seminars and in-person in-country projects, to Reach University with its apprenticeship degrees. Finally, because it is difficult to redesign existing institutions, we should encourage experimentation from both existing schools as well as new AI-native colleges and university.
That will require accreditation reform that lowers barriers to entry while strengthening the focus on student outcomes. Something counts as innovative only if it creates true progress and value for students and society. If we focus on student outcomes, AI can become not a threat to higher education, but an enabler of redesign that strengthens its value for every American. Thank you.
Burgess Owens
Thank you. Under committee rule nine, we will now question witnesses under five-minute rule. I will recognize myself for five minutes. Mr. Pozard, Florida State has partnered with companies like Google, Microsoft, Amazon Web Services to expand access to AI tools and training. How do partnerships like these bridge the gap between classroom learning and skills employers are seeking?
And how can universities structure themselves to give students workforce relevance experience while preserving institutional independence and academic priorities?
Jonathan Fozard
Thank you, Chairman. I think that this is a, this is a great question because it's a partnership. Higher education, we don't view that AI will replace instruction or replace the classroom, but the partnership will prepare, will prepare our students to be able to have exposure to the tools and the workforce and the platform that they will use in their future careers.
So for us, these partnerships with companies like Google, with Microsoft, with Amazon is to give them access, that ubiquitous access across all disciplines, across all fields to have exposure to these tools in a secure and reliable way. And our framework that we've developed is called RISE. And it's built on Research and Instruction for R, Innovation, Modernization for I, Security and Compliance for S, and Engagement and Student Success. And we really feel like our AI strategy has to be built on those four pillars to be successful with any partnership that we do. But we do feel like that there's a classroom component that we have to have foundational access, foundational knowledge, but we also have to expose our students to the tools that they're going to use in the workforce.
And so Florida State University has taken the approach of building these partnerships and having secure environments where their data is protected, they can experiment with our instructors, with our researchers in a secure environment.
Burgess Owens
Very good. Thank you, sir. Dr. Duke, you spent many years working in educational technology. The rapid rise of artificial intelligence since 2022 has transformed the conversations around teaching and learning technology. From your perspective, how has educational technology landscape changed during that time?
And what are the biggest implications of that shift for students, faculty, and institutions?
Dave Duke
Thank you for the question. It's true, the landscape has changed significantly. As we sit here today, AI has been embedded in every daily workflow for students, instructors, and administrators. The vast majority of students are using artificial intelligence. Faculty adoption is uneven, but it is accelerating.
Everyone from admissions counselors, financial aid representatives, other administrators are all using generative AI in the work that they do. Virtually every education infrastructure company like McGraw Hill has included AI-enabled products into their ecosystems. And so AI, as we sit here today, is now a requirement to provide the best learning possible. It has increased personalized learning, it is increasing efficacy, and if we don't invest more and put more resources behind this, we won't see the full value of artificial intelligence.
Speaker 10
Thank you. Mr. Horn, the rise of artificial intelligence has led some students and families to question the value of college education. What must colleges and universities do to demonstrate their value in the AI-driven economy?
Michael Horn
It's true that many people are increasingly questioning the value of college. I think what it's incumbent upon colleges to do is really show the outcomes, the connection to the labor market value that we know that families are expecting as a requirement for jumping into college. That's number one. And number two, show the relevance, the real-world connections, the programs that you just heard about at Florida State, real connections with employers will go a long way to showing that we have their interests at heart, and that's incumbent on every college to make sure that those connections are clear.
Speaker 8
Okay. Thank you so much. I now recognize the ranking member for purpose of questioning witnesses.
Alma Adams
Thank you, Mr. Chairman. Dr. Burns, as you may be aware, HBCUs play a major role in expanding educational opportunities, especially in STEM-related fields. A joint report from UNCF, Huston-Tillotson University, and Edulytic-Zen highlighted that AI could transform the role of higher education at HBCUs by creating new opportunities for innovation.
However, simultaneously, many HBCUs continue to face challenges with funding and resource constraints that can limit their efforts to access AI technologies to the extent that these schools decide to use them. And that matters because AI readiness cannot only be available to the institutions with the largest endowments, the strongest vendor relationships, or the most staff capacity. If we're not careful, AI could become one more area where well-resourced institutions move forward while under-resourced institutions are asked to do more with less. Mr. Chairman, I'd like to, to, for unanimous consent to enter this article, HBCUs, artificial intelligence, and a new vision for higher education into the record. Thank you.
Dr. Burns, what are some of the opportunities and challenges that HBCUs are encountering with AI innovation?
Bridget Burns
Thank you, Ranking Member Adams. I would say first, the UIA includes the largest HBCU in the country, North Carolina A&T, and so we have some familiarity with the challenges and they are quite relatable. We're seeing this across the sector where it's a, I would just say broadly, the lack of broader coordination across higher education to ensure that all institutions are benefiting in real time when institutions are actually figuring out some of these similar challenges around literacy and safety and privacy. That is disadvantaging HBCUs, it's disadvantaging under-resourced institutions broadly. I observe that the indirect cost challenge around the research enterprise has definitely made it very difficult for institutions to navigate.
We're asking them at a moment when we need American ingenuity, which our academic research is really the heart of, is, is operating in a time when it's been relatively destabilized, I would say, financially. But these HBCUs broadly are an incredible STEM talent activator. They are, they generate incredible talent and we just believe that broadly for the competitiveness of our country that we need to activate all the talent we have.
Alma Adams
Thank you. And thank you for mentioning North Carolina A&T, that is my alma mater twice. Dr. Burns, colleges and universities now have access to many different AI tools. Some come with packages of other technology such as Microsoft 365, while others require specific subscriptions. Given that larger institutions have more resources for new tools, under-resourced institutions may have a harder time providing these tools and developing faculty understanding, therefore creating disparities in our students' academic careers.
And I wonder that if, if students at one institution are receiving strong AI literacy, hands-on experience, and access to tools employers are already using, while students at another institution are not, then we're not preparing students equally for the future of work. So given that larger institutions are at the forefront of developing AI curricula and competency, what impact do you think that this will have on student access and the digital achievement gap?
Bridget Burns
I am deeply concerned about students who are being left behind across the country. We noticed early on, you know, it's, it's not like we have a ton of expertise with navigating AI, right? Everyone is as new as a couple years at this technology. And so when my institutions were first trying to diagnose how they decide to sign any of these agreements, ASU, for instance, was the first institution to sign an enterprise-level contract with OpenAI. We started in that moment sharing what are the questions you ask, how do you interrogate it, because this expertise doesn't exist.
Normally, it wouldn't be just like the legal counsel knows how to navigate something so complex. And it's worth definitely seeing it at HBCUs and there are some resources out there, but I think there need to be more.
Alma Adams
Thank you. A few more seconds. What can under-resourced institutions do to provide additional access to these tools despite limited capacity?
Bridget Burns
Ranking Member Adams, I believe that institutions need to be broadly sharing what they're learning in real time. We currently lack the coordinating infrastructure where one campus is tinkering and others are going to ask the same question and it's actually going to harm students. I'm most concerned about institutions that are in a moment where they don't have that capacity, they will freeze. That risk is too, we're too uncertain and we can't delegate the responsibilities to our student to a vendor. So as a result, there are some institutions who will resist, who will just freeze because they are waiting for more clear direction.
Thank you very much, Mr. Chairman, I yield back my amount of time.
Speaker 24
Thank you. I'd now like to recognize the chairman of the full committee, my good friend from Michigan, Mr. Walberg.
Tim Walberg
Thank you, Mr. Chairman, and thanks to the panel. And I wish there had been a miracle today and all of the committees that are meeting would have not met except this one. Because frankly, it would have been great and Mr. Chairman, thanks for your wisdom in holding this, this hearing. It is so important, but I think that if we all had been here on either side of the aisle, there'd be a great amount of discomfort on our faces at what we are hearing because of the challenges that we are forced to face. But Ms. Burns, thank you for giving us some examples of hope and opportunity.
And also the opportunity for me to say, this is a time of disruption. And higher education needs to find ways of doing things without always more money, but being creative and setting priorities. So I appreciate the examples and again, I wish we all had had a chance to hear because these are great times, but the challenge is there and we don't want to miss it. And Mr. Horn, I was taken by your example of electricity and taking that wonderful thing and and just doing the same old same old until finally we woke up. And I think that'll be the case with AI.
But the Chinese Communist Party isn't waiting for us. And we had better be the ones that are using it well. So could you go on a little bit further, Mr. Horn, and or Dr. Horn, and tell us what it would mean for a college to redesign itself around AI rather than simply adding AI as a tool?
Michael Horn
Absolutely. On our research on innovation, what's become clear is that institutions need freedom to free up the how they do their work to be able to actually rethink the resources, processes, the financial formula in some cases to your point about doing things with less and really reinvent themselves. And that requires unconstrained in terms of the inputs and what we're looking for in the institution, but an absolute rigor on what the outcomes are on the other side. And I believe what we would see and we are seeing in many pockets already is a plurality of different designs where we know, for example, Western Governors University is already creating a native AI university experiment. We know that UMass Global is doing the same thing right now.
And so we're seeing existing institutions literally carve out separate areas where they can rethink the fundamental first principles of what these universities look like. And frankly, we've done this before with online education as well. Many institutions just used online education to layer over their existing models, but some institutions, again like Western Governors, started thinking from first principles and they have a very different model as you know with taking the faculty role and unbundling it into five different roles to create a much more robust web of support, but also a very affordable model for students that has a very high success rate.
Speaker 16
Again, and again an example of where the consortiums and joining together and sharing this is so important.
Speaker 14
Mr. Fazard, how should universities respond when students are making decisions based on uncertainty about AI's effect on future employment?
Jonathan Fozard
I think that's a question that all universities should be asking. We're having these conversations with our leadership, our provost, our faculty are having these. I think that it is every discipline, no matter what discipline it is, will need some type of AI literacy and AI skill set. There will be degrees and programs that we will have to think through the curriculum and we're going to have to evaluate what those programs look like. But I strongly believe that every discipline, every field, every degree that we have will need some component of artificial intelligence baseline literacy and use.
Just plan on it.
Speaker 17
Dr. Duke, what kinds of evidence do you look for before concluding that an AI tool is truly improving learning, and how do you ensure those tools are helping students build lasting understanding rather than completing assignments?
Dave Duke
Thank you for the question. This is something we do pretty well, efficacy research, and we approach this with regards to GenAI in a similar way in terms of backwards design. So we do pre- and post-test measures, basic statistical analysis. Said differently, we have an idea in mind what learning outcome we want to create and we work backwards: did we create a learning experience that produced that? The new layer, though, with generative AI is who did the work?
So when GenAI is involved, we have to make sure we have preserved the productive struggle of students so that we can assess whether or not they did the work and therefore we had the outcome we wanted.
Speaker 17
Well, thanks again for scaring me, but also exciting me that my grandkids are going to have an interesting life ahead because of things that we can use and learn from screwing up with electricity originally. Move on. I yield back.
Burgess Owens
Thank you. I'd like now to recognize my friend from California, Mr. Takano.
Mark Takano
Thank you, Mr. Chairman. Dr. Burns, thank you for mentioning the university in my district, the University of California, Riverside, in your testimony. As you point out, instructors at UCR are using AI to prompt students to learn how to be better writers rather than to do their work for them, that using AI to kind of coast or to make it look like they wrote a product, but they're actually using AI to teach the students to actually be better writers. Innovations like this one are one of the many reasons why UCR is the top university in the country for social mobility. And what I mean by that is that students of lower economic means come to this university and they gain the capacity to actually increase their economic situation.
And there are measures out there on social mobility and UCR is one of the universities where the students who go there make incredible gains. Now, innovations like the one we just mentioned are one of the... so, Dr. Burns, what I want to know is, do you foresee more examples like the one you gave in which AI technology could lessen the gaps in student achievement?
Bridget Burns
Congressman Takano, yes, in fact, I had a difficult time winnowing my testimony greatly because there were so many examples across our institutions. There is so much creativity and innovation that has been unleashed by this technology. I was just on UCR's campus two weeks ago and you would be blown away at the incredible way that we are unleashing the creativity of faculty to reimagine how they can support students. And really for us, when you think about closing gaps and thinking about how you can help all students, the, you know, the holy grail we're looking for here is personalized learning. And this is going to enable us to possibly achieve personalized learning at scale with the small pilots and experiments that faculty are doing in real time.
And, and we are measuring not just by the degree outcomes, but rather we're moving to that post-college mobility as being the measure of our success, something that UCR has been a leader on.
Mark Takano
So as a former teacher at the secondary level, and having been a substitute teacher at the primary level, the individualized learning is what I see as the greatest potential of AI. Not replacing the teacher. You'll need teachers for that human interaction and that human connection. But a teacher with 30 kids or 180 kids at the higher level, but it doesn't change much from like 12th grade to like freshman year when students have got to pass usually the expository writing requirement and the university math, university-level math. And one of the things that I as a former community college trustee that I encountered a lot, what was an ongoing theme is, is the cost of remediation when students either, you know, they enter straight into the four-year university or that it's a struggle for the community colleges to get students to not, to retain them.
What, what have you, what's, what sort of things are going on in sort of addressing this remediation? Because I kind of think of that as one of the gaps. Would you agree with me?
Bridget Burns
Yes, I think this is one of the things where you really need as much of the capacity and technology is a, is a really supportive tool when, when used well, to ensure that we can understand what a student knows, what they don't know. Otherwise, it's very difficult if you're teaching in a course to actually know where every student is coming in from. So this is definitely one of the ways we're thinking about it. We just wrapped up a complex project to try and address D, F, and withdrawal rates on campus, especially in gateway courses. And we applied an intervention that made a huge difference, adding supplemental instruction and near-peer coaching so that students actually had that support.
And some campuses are experimenting with using AI as a coach, as actually using that Socratic method to be able to probe students. It's actually helping with learning. And that's where we need more studies is to understand how we're going to actually be able to measure the learning improvement. But we already saw that it increase, normally when a student retakes a course if they've taken, if they've gotten a D or an F in the past, they've only passed at about 50% that second time. Ours were up to 80% they were passing the course.
So we're seeing incredible gains. There are lots of experiments happening. I would say that our institutions are, what really matters is not just the number of experiments, but how do you make sure that the ideas are spreading? The scale of learning and how we accelerate ideas across the sector so that every institution can benefit from the experiments and the innovation that are underway.
Speaker 13
Well thank you. I wish we could have spent more time together, but I got to be respectful. I yield back.
Mark Harris
Thank you. I would like to recognize my friend from North Carolina, Mr. Harris.
Thank you, Mr. Chairman, and thank you to all of you on the panel today for your expertise. And Mr. Chairman, appreciate you calling this, this hearing. You know, over the last several years, we've seen a dramatic increase obviously in the use of AI programs. And this increase has been apparent in academic settings from K-12 institutions to secondary schools to colleges and universities. And while many students have used AI to complete coursework, there has been a need across the workforce for students to have more of an in-depth knowledge of how to use AI technology in professional settings, as you all have, have touched on today.
Closing that knowledge gap between students using AI for coursework and the young professionals entering the workforce where proficiency with AI programs is necessary for creating work products is absolutely critical.
Speaker 12
And Dr. Duke, I just noticed in reading your written testimony you explain the ability to work with AI tools productively, critically, and responsibly within a specific skill domain is an expectation that is shown in job descriptions for entry-level jobs across the country. I would ask how can we help prepare students to transfer their AI knowledge from basic AI proficiency learned in school to the skills that our workforce are actually looking for.
Dave Duke
Thank you for the question. I think one of the missteps we can make is to synonymize enabling a workforce with artificial intelligence and applying AI to the learning journey. They're not the same thing. In the workforce you want people to be very, very efficient, work efficiently, make very few mistakes. When people are learning, you actually want that to be inefficient sometimes, you want them to make mistakes, that's where the best learning happens.
And there is the dilemma. Along the way, these students have to establish proficiency while also managing learning. And so one way we can do that is to integrate it into the curriculum. We could define the set of skills that students need and make sure that they get them along the way. They get some of this proficiency through daily use, but it's not the same as cultivated professional competency.
And this is where I think the learning science actually needs to catch up and how we need to change concept mastery and how a student marches toward not knowing anything and graduating and being in the workforce.
Mark Harris
And you also discussed the importance of treating AI proficiency as a quote core educational competence just like information literacy, quantitative reasoning, or financial literacy. What would you in your opinion treating AI proficiency as a core educational competence look like in the higher education setting?
Dave Duke
Well one we would have to say that, we would have to determine that it is a core competency or something that is expected. Right now I don't think that we do that. We do that in other areas like financial literacy and information literacy as you point out. But until we do that, institutions aren't going to be able to design their curricula around that. And also making it a core competency I think will accelerate some of the thinking, the experiments, and the testing that we've talked about on the panel today and will help us improve in this area.
Mark Harris
Thank you. On January 23rd, 2025, President Trump issued an executive order titled Removing Barriers to American Leadership in Artificial Intelligence. And although that EO does not specifically address higher education, the goal of the order was to revoke restrictive policies and promote innovation. And Mr. Vizard, I just want to ask in your experience working at Florida State University, what impact does removing restrictive policies and promoting innovation have on the way AI is developed and utilized on college and university campuses like your own?
Jonathan Fozard
I think that artificial intelligence has to have the ability to be integrated into curriculum, be accessible to our students. A case in an example is we partnered on a design sprint. It was a 24-hour design sprint with students across multiple disciplines to solve a real-world problem. AI didn't do the job for them. They had to come together, they had to collaborate, they had to use AI in these in these new and innovative ways to solve a problem.
So there was a problem, you get a team together, they have 24 hours to work around the clock to solve it. It's something that they're going to do in the real-world setting too in their careers. And I think that we have to allow the space for innovation to take place. The other thing that I would say is we have been heavily investing in the material science space, the engineering space, the quantum technology space at Florida State University. And that's under the leadership of our president because that will differentiate us too to have that infrastructure.
And we must be able to research that and to invest in that to be able to really be a leader in the world in artificial intelligence.
Suzanne Bonamici
Thank you, Mr. Chairman. Thank you to the witnesses. Dr. Burns, thank you for all the work you've done in higher education in Oregon and now, of course, with the Alliance. AI, we know, is shaping classrooms and workplaces faster than many schools and employers can respond. I'm also on the Early Childhood K-12 Subcommittee and we're having similar conversations there.
But I'll tell you, many educators at all levels are concerned about a decline in their students' creativity and critical thinking skills. Dr. Rebecca Winthrop from the Brookings Institution calls this cognitive stunting. And if it's happening at the K-12 level, it's definitely going to be affecting higher education. It's the impairment of cognitive development that results from an over-reliance on AI tools in the learning process. I'd like to submit for the record a recent article by Dr. Winthrop that's in the New York Times titled "What 370,000 College Essays Tell Us About AI's Effects on Creativity."
Dr. Winthrop discusses the importance of the brainstorming process for building creative thinking skills and generating new and innovative ideas. And offloading the act of brainstorming is one of the many ways that, despite its numerous promises, AI can do more harm than good for students in the long run. So I'd like to submit that article for the record, Mr. Chairman.
No objection.
Suzanne Bonamici
And UC Berkeley School of Law has a new policy restricting the use of AI in academic work submitted for credit. They want students to use their cognitive skills. The policy prohibits AI in conceptualization, revision, translation, and exams. A recent article about the California State system just out discussed a survey that a majority of students and faculty are skeptical of the benefits and worry about what AI will do to not only job security and creativity, but also the environment. And there's a lot of questions about AI grading.
I'm a parent who paid college tuition for two students for many years. When students are paying tuition, do they not deserve human feedback? And so there's questions about AI grading as well, privacy, security, bias. So I just want to put that out there. As AI continues to integrate in higher education, I think it's more important that institutions put appropriate safeguards in place to protect students from the risks, including from the misuse and harmful effects, and protect against bias, protect privacy and security, and also the decline of critical thinking skills.
So Dr. Burns, what safeguards should be in place to protect students from the risks associated with AI, including cognitive stunting?
Bridget Burns
Thank you, Congresswoman Bonamici. I share your concerns. These are things that our institutions are wrestling with right now in real time. Every institution in the country is. And again, I'm most concerned about under-resourced institutions navigating this because these are things that are absolutely on the frontier of science.
We're trying to figure out the learning effects, the mental effects, the cognitive load. I would say that you want more testing, you want more validation, and you want a thoughtful process where institutions are evaluating materials. But I think in terms of privacy, we know that and the protections for students that we absolutely are responsible. And we cannot delegate that responsibility for student privacy to a vendor, which is where you create the problem because...
Suzanne Bonamici
Right, right. And I hope that responsibility is shared among all institutions of higher education. The World Economic Forum's Future of Work report just forecasted that AI literacy, creative thinking, and curiosity will all be core skills in 2020. But AI literacy involves more than just learning how to write prompts.
People need to know how to critically evaluate AI through ethical and linguistic and social perspectives that reinforce that critical thinking. I'm thinking of the example of the when we we recently spoke with Dr. Winthrop, she was talking about how like a 10th grade student will write a paper using AI and then put it through a humanizer to make it look like a 10th grade student wrote it, so they take out sophisticated language and put in some typos. And then if if then the AI is grading it, I mean, what what are we doing in education? So what is the current state of AI literacy in higher education, and how are institutions integrating AI literacy across fields of study, especially among the humanities, which remain important?
So our institutions, like I shared, Purdue already has put this into their graduation requirements, is AI literacy. Other institutions are following. We believe that it is technical fluency, so you need an understand how it works and why, when to use it, but it also has to have those durable human skills as the as the back end. You need the critical thinking and the judgment to work alongside AI.
Suzanne Bonamici
Absolutely. And and I've mentioned this before, I'm working on a comprehensive human-based framework to to strengthen teaching, learning, and dignity of work in an AI-enabled economy. I look forward to working with my colleagues, hopefully on both sides of the aisle, truly promote that critical thinking, which will remain important going forward to improve literacy and professional development, which there is not enough of in not just in the higher ed system, but also in the K-12 system, and also protect privacy, safety, and civil rights. So thank you very much, Mr. Chairman. Thank you to the witnesses, and I yield back.
Robert Onder
Thank you, Mr. Chairman. And thank you to the witnesses for coming before the subcommittee today. It's no secret that our our students are increasingly outsourcing their cognition to generative AI platforms like ChatGPT. An October 2025 nationwide survey of college students showed that 64% of students used AI to get help with coursework and 36% used AI to write their papers. There is concern that student outcomes are worsening because AI thinking is replacing student thinking.
A September 2025 survey of employers revealed that only 22% of hiring managers believed that entry-level employees were prepared to do their jobs. The reason, according to most of these employers, is that the upcoming generation lacked the skills needed to thrive in the workplace because their heads were buried in their cell phones. Constant phone usage decreases attention spans and leads to poor academic performance. A study a study monitoring the cell phone usage of university students in Brazil showed that for every 100 minutes of daily cell phone usage, there was a corresponding drop of 6.3% in academic ranking, with with usage during class nearly doubling that drop. And when young people are simultaneously addicted to technology and performing poorly in class, they search for an easy way out, and they may use their AI chatbot to write their papers for them or give them answers to test questions.
They can outsource critical thinking skills instead of working through problems or asking teachers or peers for the help that they may need. I commend institutions like UC Berkeley School of Law that adopted an AI policy prohibiting students from using AI except to identify sources for research papers. I think it's worth reading a few sentences of that policy aloud. Quote, future lawyers may need to use AI fluently, but the current state of technology requires AI use to be coupled with the cognitive skills necessary to strategically deploy the technology, to critically assess its work product, and to uphold ethical obligations to clients and to the legal system.
In short, thinking remains the sine qua non of good lawyering and of a quality legal education. This policy seeks to ensure our courses focus on requisite cognitive skills. It provides students with the opportunity to develop their skills that they need to conceptualize, outline, draft, revise, and edit their work by forbidding use of AI for purposes in connection with work submitted for credit. This policy was crafted specifically for legal education, but I think it can and should be applied in other academic settings. Mr. Horn, what policies do you think institutions of higher education can adopt applicable to any academic discipline that ensure that students are not outsourcing their thinking to AI?
Michael Horn
Yeah, first I want to say I appreciate the point that you have made and the concern. I think it's a very real one at the current moment, and I think it speaks to the importance of building foundational knowledge.
Robert Onder
And I said a lot about cell phone use, but I think AI has the potential of these harms we see from cell phone use on steroids.
Michael Horn
And what I would say is this is where we need nuance in the conversation, right? 100%. We need, as opposed to an all or nothing, intentional use where it's supporting the kind of personalization we heard earlier, great. When it's to do the cognitive automation where you're not doing the work, that's a huge problem. And so I think building knowledge, foundational knowledge so you can do the critical thinking and judgment and understand when it's going wrong and right is critical, and having a clear set of what's the learning objective, what are the tools allowed for it, how are we going to assess it.
Often that should be real-time oral defenses, real-time conversations with faculty where you can't rely on the crutch. But then conversely, I know many recent graduates, for example, I was recently talking to someone who produces graduates in film studies. They were applying to entry-level jobs that were expecting them to use AI tools in their production and they had no experience with it from their education. That's a problem on the flip side.
Speaker 20
Thank you. Mr. Fazard, does Florida State have any plans or recommendations in this area?
Jonathan Fozard
So we freely and openly share, we actually have guidelines that are posted that we worked with our faculty, our students, our university leadership to post across all disciplines. So we freely and readily share. We also have three K through 12 kind of laboratory charter schools. So we're not just a higher education institution, we're looking at this through the whole lens of the K through 12 all the way into higher education and then those that enter the workforce after graduation.
Thank you. I yield back.
Thank you. I'd like to recognize my friend from Arizona, Ms. Grijalva.
Raúl Grijalva
Thank you, Mr. Chairman. As the representative for Arizona's 7th Congressional District serving the University of Arizona, Pima Community College, and a diverse student body, I believe that innovation is only progress if it expands opportunity for everyone. Crucially, we must recognize that if we do not provide proper guardrails, AI can quickly turn from a tool for success into a weapon used against our most vulnerable students. I am concerned that while institutions are being pushed to adopt these new tools, the current administration is dismantling the very federal offices meant to protect students, including the Office of Educational Technology and the Office of Civil Rights. These are entities responsible for providing the guidance schools need to prevent documented harms, and also as a school board member for 20 years, these offices were very critically important to the work we did at a very local level.
And like the Stanford University study showing AI tools falsely flagging non-native English speakers for plagiarism. This shows that these risks are not rhetorical, they're happening right now. We cannot ask our colleges to be AI-ready while the federal government is stripping away the oversight meant to ensure equity and justice in the classroom.
Dr. Burns, it's a pleasure to have you here, especially given your ties to our state's higher education system through your time at Arizona State University. Given your leadership scaling student success across the country, I'd like to start with practical ways we're seeing these technologies implemented on our campuses. Colleges and universities are navigating how to integrate AI into academic and student support functions, including academic advising and other student success supports. Reports indicate that some colleges and universities are seeking out strategies to integrate AI into curricula and instruction to advance student success. Could you share some examples of how some colleges and universities are scaling student success initiatives and support through AI?
Bridget Burns
Thank you, Congresswoman. Yes, we actually, this is one of the challenges as there is so much happening in this space. I shared earlier about one example where the writing faculty, which was the first department that was most impacted by AI, initially it was the writing departments. They started seeing it when ChatGPT was first released. They had to immediately figure out how you're going to adapt and how do we make it so that this is not going to address some of that cognitive burden challenge.
So figuring out how you use tools as a Socratic learning tool, how you actually can create ways that it's a prompt that's actually stimulating more creativity and more learning from students. That's happening in real time across all of my institutions. And I would say the number of experiments is quite significant and they are sharing in real time. Part of the challenge is they're literally using AI in so many different facets of the institution and trying to find and develop the coordinating infrastructure that brings those different entities together, which did not previously exist, is part of the complexity in the moment.
Raúl Grijalva
So building on those concerns, I want to specifically look specifically at the accountability vacuum left by the loss of federal oversight I mentioned earlier. Since the federal government has eliminated the offices that once provided safety toolkits and civil rights guidance, we're often leaving institutions and students to fend for themselves when technology fails. With the loss of these centralized federal oversight bodies and the guidance they provided, what entity is currently responsible for holding institutions or private tech vendors accountable when their AI systems produce these types of discriminatory results? And that's for Dr. Burns.
Bridget Burns
Congresswoman, what we are, we're very clear that this is actually something that is the responsibility of our own institutions. We cannot actually delegate the responsibility for when it comes to students' privacy protections, which is part of where the challenge is, is that we're having to evaluate every single vendor individually. In general, there need to be privacy protections that are very clear, enforceable, and modern. And no matter where that enforcement is happening, I know there is transition currently, we just need that, I would say that institutions need the capacity to evaluate the tools and someone needs to provide that kind of enforcement mechanism regardless of where they're located. They need to be able to help guide institutions as they're navigating the constantly changing environment.
Raúl Grijalva
Yeah, thank you. Beyond just identifying these risks, what specific pathways for recourse currently exist for a student who's been unfairly penalized or academically disciplined due to bias algorithm?
Congresswoman, again, we are very clear that this is not something that we can delegate the responsibility for enforcement to any company. And so we take that responsibility very seriously and we believe very strongly that a human must be in the loop on any consequential decisions. And so in any of the examples I'm sharing, if a decision is made, there has to be a human brought in and that's how we're building from the ground up.
Okay.
Glenn Grothman
Thank you. Now I'd like to recognize my friend from Wisconsin, Mr. Grothman.
Thank you. We'll start with Mr. Horn. You emphasized that students need real-world work experience because AI is rapidly changing the nature of work. How can internships, apprenticeships, and employer-connected projects help ensure students graduate with the necessary skills and experience?
Michael Horn
I think when you think about building student understanding of what the workplace will look like, there's three modes to think about: is learning about work, learning through work, and then learning at work. And so a college program that starts in the early years with bringing exposure to students so they can understand, hey, this is something I actually want to do, this is something I don't want to do and rule it out, that iterative nature is important. Secondly, when they're actually interacting with employers in these settings through projects, apprenticeships, internships and the like, they're actually building real-world connections, real social capital, real mentors that can help guide them and get them into jobs in the future. One of the myths I think that the younger generation often has is that you get a job by applying online, when research shows upwards of 50%, maybe as high as 85% of jobs are filled by someone in your network. And then thirdly, you actually get to experience the real ways that technology is used in these jobs and what professionalism and the durable skills that you need to bring to bear look like in these roles, and you build that fluency while you're in school.
That's the entry-level experience you need to get that first job.
Glenn Grothman
Okay. What barriers do you think students face when trying to just start out with an internship or apprenticeship related to AI?
Michael Horn
I think the biggest thing right now is that schools are not providing the infrastructure nearly enough to actually make that a default option. What I hear from a lot of the schools that I talk to is, oh, we offer that, but only our top students take advantage of it. This should be the default for every single student so no one's falling through the cracks. They should be doing it on the front end of the academic planning so it's part of the process. Career services shouldn't be the afterthought that's at the end of the process.
Speaker 30
Okay. Mr. Fazerd, one of my concerns about AI is that it could lead to a decrease in human involvement in the workplace due to over-involvement on technology. How is Florida State's Information Technology Services internship program ensuring that human oversight is not removed?
Jonathan Fozard
Thank you for that question. I actually was a product of a student internship. I got into my career in technology because of that. And we have 150 student interns that go through each semester, hands-on experience, hands-on learning with artificial intelligence and how we can use it at the university. And those are positions from cybersecurity to helping assisting with our enrollment to doing student support services.
And our results in those internships show because we have a 99% fall to spring retention rate in our university. And we're using these students not just as workers that are grabbing coffee or making copies, but actually using the technology and having hands-on experience before they graduate. We're seeing these students not just succeed, but thrive in their careers.
Glenn Grothman
Okay. As a broader question, one of the things that concerns me about AI, when you look at other platforms even in the past, even if you just Google something, when you begin to look over time, you see an ideological bent to what comes back. And as it is, you know, you read articles generated by AI and over time you begin to see a little bits of ideological bias as well, in which maybe things in history disappear and things that were minor in history all of a sudden are exaggerated. It becomes a very, very scary thing. And I don't know what to do about it, but would any of you like to take a shot on what we can do to make sure in the future I can trust if I use AI that I'm getting something that is factual, not ideologically tilted?
Jonathan Fozard
Just briefly, I think that this is where education has to play a major role.
The critical thinking skills cannot be replaced by artificial intelligence. And this is where K-12 education and higher education must be present and we must equip our students with those critical thinking skills, um, to be able to leverage those tools and be able to question those tools.
Glenn Grothman
But my question is, I think people, what they learn is going to be based on what AI gives them. Right? We already have the schools, I believe, overwhelmingly biased because we have overwhelmingly biased faculty. Now if you have overwhelmingly biased AI, or even a little bit biased AI, in which irrelevant parts of history are, you know, you can, you can create a world view. I don't see how we do that, no matter how good the student is, if his reference becomes AI.
Jonathan Fozard
I, I would just argue that, um, you know, any technology, any media, anything that we have access to, we have to use critical thinking skills to be able to determine and to differentiate fact from fiction. And so I view AI as a technology and a tool. It's not a, a magic eight ball. Um, it has to use data and it has to have, um, compute power to be able to give and provide answers. And I think that if we view AI through this lens of, um, being kind of the sole source of truth, we negate the human aspect, which I think is so, so important to be able to differentiate fact from fiction.
Speaker 31
Okay, thank you. Thank you. Now I'd like to, oops, now I'd like to recognize my friend from California, um, Mr. DeSaulnier.
Mark DeSaulnier
Thank you, Mr. Chairman. Um, well I want to thank you all for being here. It's fascinating. Um, I'd like to talk first of all to Mr. Burns about, um, the behavioral health aspect of kids as they develop and go to college. I think, uh, where am I, sorry, Dr. Burns.
Um, I just flew in, so forgive me. I'm jet lagged. Uh, I, speaking of jet lag, so the University of California, I think recently over the weekend was talking about reestablishing standardized testing because so many of the students that they're getting, the, the professors just don't feel like they're noticeably not prepared. Um, so when it comes to behavioral health in kids, we've got a lot of good research now developing about the effect on cognitive development in social media in general, um, but particularly with AI, it's quite frightening. I just had a meeting with the authors of the AI Con and, and it's just, uh, it's a combination of marketing and as a member from the Bay Area who has friends in tech who warn me that, uh, we're doing the same thing with AI that we did with social media, we're marketing.
Um, and unfortunately, caveat emptor, I don't think kids are prepared to be aware of, particularly the sophistication of how their brains are being developed when they're on, um, on this AI. So could you speak a little bit about that as first do no harm? How are we dealing with cognitive development and AI and how do we get at least a, a standard of first do no harm before we really know how bad this is, particularly by the time they get to higher education?
Bridget Burns
Congressman, I, uh, I think your concerns are shared broadly, um, not just within the academic environment where we're seeing that this is going to create long-term effects because this is a workforce readiness challenge as well. Um, I think this is a place where we need more investment in federal research because we need to have far more testing, uh, to happen not just, you know, before, like on the actual models as they're developing, but the, the short, medium, and long-term effects on students. It is happening so quickly and, um, I would say this is a space where there is a, a need for not just, uh, federal research investment, but to have it be multi-institutional so that we're actually pulling together the expertise across the country because I think this is not something you want one institution or one entity to figure out how we do it better.
You want this to actually spread broadly.
Mark DeSaulnier
And I have a bill just that I have with some other members just it's like a warning bill that would just tell parents that when you go online, this is just for social media, but to make it stronger for AI. Just sort of like tobacco or liquor, make parents and kids be aware that when you go on here, you really don't know the power and how sophisticated it is. Could you respond to that and then anyone else who would respond to it? And the just the gist of everything I'm trying to get at is having spent a lot of time on this, having spent a lot of time with experts in the Bay Area at Stanford in particular, both neurological pediatricians and people in the education school there is and their relationship with artificial intelligence and with technology. How do we have a discussion about we don't want to inhibit innovation, but on the other hand, we've got to slow it down so we make sure the full consequences and the fact that most people don't realize how targeted these algorithms are becoming.
Bridget Burns
Congressman, again, I just want to reinforce that your your concerns are shared broadly for parents everywhere, young people, and our faculty are contemplating this on a daily basis as students are coming into the classroom. We're trying to analyze exactly what is happening and to really study the learning effects. I would say that this field is somewhat nascent, that this is a space where we need to do more research. We need to actually interrogate this all along the way. But again, it also is just so clear to me the lack of broader coordination to ensure that institutions are pulling together on this.
This is not a challenge that any one human being will be able to to tackle or one institution will tackle. You need the entire higher education sector to be enrolled in figuring out how to use AI responsibly because we take protecting our students very seriously and we don't delegate that responsibility to anyone else. The real thing is who's missing up here are the AI providers. And that's these models are being released so quickly and the question of whether or not they're having the level of efficacy testing prior to release is something that definitely needs to be paid attention to.
Mark DeSaulnier
No, and I hear that in the Bay Area when I go out and talk to people. They're like, oh, you're inhibiting innovation. I said, that's what you told me 15 years ago. And I don't want to inhibit innovation, but we've got to be careful of the consequences. You can make a lot of money, but we're not fully aware.
And one of the best stories on this, and I'll wrap up, sorry I didn't get to have you respond, is the YouTube story about pre-adolescent, adolescent girls being targeted by their algorithm so that 30% of them were predisposed to suicidal ideation and depression to the point where the YouTube algorithms would bring them all the way to a YouTube that showed them how to commit suicide. And if it wasn't for a reporter, we might not have found that out at all. So with that, I will thank you. I apologize for not having being given you enough time to talk answer my questions. And thank you, Mr. Chairman.
I yield back.
Randy Fine
Thank you. I'd now like to recognize my friend from Florida, Mr. Fine.
And I've thought about how his high school and junior high career has changed as he gets ready to go to college. My my oldest son graduated from high school two weeks ago. Well, thank you, Mr. Chairman. For the last couple weeks, I've been having a lot of nostalgia. He was one of the kids who went home for spring break in sixth grade and then had the longest spring break ever because they never went back to school because of COVID.
And then his junior high was ruined because of many of the terrible decisions that government made about how to deal with COVID, something that the kids graduating now about to go off to universities are probably behind forever because of the decisions that were made. But now I've gotten to watch over the last year how AI has affected his senior year. And fortunately, he was smart enough to realize that if you had AI write your paper, there might be AI that could figure out that you used AI to write your paper, as some of his classmates learned the hard way.
So we have seen that these technologies are changing how the world works. And I apologize to the other three witnesses, my questions are going to be for the home state, Florida State, where my son did apply and was admitted, but I'm sorry, he's going out of state. I won't say where. But I wanted to ask you though a question about a bill I've introduced, HR 8747, the K-12 AI Literacy and Readiness Act. So the bill recognizes that AI is becoming as foundational as computers and the internet.
It fundamentally modernizes schools to help integrate this new technology, equip teachers, and prepare students to compete in an AI-driven economy in a safe and responsible way. Do you believe that if K-12 schools begin preparing students with AI concepts and tools, will those Florida students be better prepared to succeed at places like Florida State and in the economy shaped by AI?
Jonathan Fozard
Congressman, I think that you have to have those baseline skills in the K through 12 space to translate into higher education and translate into the workplace. I think the earlier that we can get those skills in a safe, secure environment with our students, the more likely that they'll be successful in using the technology. And to the other congressman's question, they'll be able to evaluate it in the right ways on how they're using it with the critical thinking skills and in what works and what doesn't.
Randy Fine
And my bill, to be clear, isn't just about training students. Students to some degree can figure it out faster than we can because they're younger. My son is much more advanced in the use of AI than his 52-year-old father, who by the way spent 22 years building software companies. So I see how fast the world is changing. But I want to talk about what all this means.
People worry that AI is going to replace the jobs on which we rely and depend on. Frankly, for many of I think particularly some of my colleagues, I think the people who used to write their speeches are replaced by AI. Fortunately not me. But without exaggeration, whether we welcome it or not, we should welcome this inevitable reality. Technology has changed how the world works since the horse and buggy, and we didn't look back, we looked forward.
And we have to do that here as well. Our education system though has a challenge to prepare them for the reality, to not train them for the jobs that will no longer exist, but to train them for the ones that will. So what can universities do? What could Florida State do, the other universities in our number one in America state university system in Florida? What can we do to ensure students don't lose the critical thinking, the problem solving, the communication, and other foundational skills that many believe could be eroded by AI?
Jonathan Fozard
I think it comes down to access. We have to give ubiquitous access across all disciplines, not just one discipline. It's not just computer science majors. It has to be every major that goes through a collegiate system has to have access to the technologies in a secure environment. I would also offer that Florida State University, we are giving an AI kind of foundational certification that every student has access to take to build those skill sets up.
And it's regardless of discipline, regardless of program, everybody has access to that to build those skills. I think it's very important to lead, not follow.
Michael Horn
All right, well since I have a few minutes left, I'm going to let someone from my alma mater actually answer the question, even though he's not from Florida. Mr. Horn, I don't know if you've got any thoughts on these questions.
I think this is why it's so critical to frankly build knowledge, right, as we start the progress up and recognize that the use of AI in different domains is going to look differently in different places. And so each domain is going to have to be very intentional about how it uses AI and what are the foundational knowledge building blocks and then what are the skills of integrating AI. The one analogy I would say is when we had calculators introduced several decades ago, we still learned how to add, subtract, do multiplication, and frankly memorize multiplication tables. Then you could use the calculators to help you.
As I've had to remind my 14-year-old.
Michael Horn
And it's an important fact that we still need to build those cognitive literacies.
Randy Fine
Well thank you. Look, I think the key thing is this is going to change incredibly fast. It's not just that the world is changing, the rate at which it is changing is changing faster, and we've all got to be prepared to deal with that to make sure our kids are prepared for the future. Thank you, Mr. Chairman, and I yield back.
Speaker 33
Thank you. Thank you Mr. Chairman. Ms. Burns, you indicated, Dr. Burns, you indicated that well-resourced institutions need to disseminate the information they gain to lesser-resourced institutions. Exactly how would they do that?
Bridget Burns
Congressman, thank you for the question. So this is my biggest concern is that it shouldn't be left on the responsibility of the institutions themselves. They can make decisions about how they build things. The example I shared from ASU of they actually built the entire triangulator outside of ASU's system so others could use it. The real question is who is or what infrastructure is missing in terms of ensuring coordination that institutions across the country are able to know what experiments are happening, what have you learned, can I borrow that language that you used in that contract, is this a good vendor to use or not?
All of these questions right now are lying on just people happening to know who to talk to and given that at AI, who's the actual professional responsible for AI is adapting in real time, this is a huge space that has challenges. K-12 has something called the EdSafe Alliance, which has played a critical coordinating function so that lessons learned about literacy, safety, and privacy are spreading. We think that something similar needs to exist in higher ed.
Speaker 33
Thank you. Mr. Facade, how is student data protected and not misused and how do you make sure that students know how their data is being collected?
Jonathan Fozard
That's a great question and thank you for asking it. One of the frameworks that we have at Florida State University is to make sure that we sign agreements with companies to make sure that our data isn't used to train models. It's one of the foundational aspects at Florida State University and I think every university should ensure that, that there's not free models being used that go and train, their intellectual property isn't being given away, their research isn't being given away to train a model. And so we enter into agreements and sign agreements with companies before we offer this at the enterprise level, again to every student, every faculty and staff, to ensure that our data is in a secure and protected environment.
Speaker 37
And AI can be used for administration in colleges as well as the teaching. How would it be used in admissions and administration and particularly how would you avoid racial discrimination in this application?
Could you repeat? I'm sorry.
Jonathan Fozard
Yeah, we use it across... ...and administration and how would you make sure that racial discrimination isn't embedded in those algorithms? So we still believe strongly in a human in the loop and I think it needs to be a founding principle for all artificial intelligence, that a human is in the loop in automation processes, in reviews, there still needs to be a human in the loop. And that's one of our foundation principles that we have at Florida State.
Speaker 42
So there'd be no loophole for racial discrimination just because you're using AI?
Speaker 39
The candid answer is it's uneven across different faculty and classes at the moment and we're all trying to figure it out right now in different ways.
Speaker 33
What kind of jobs do you see AI-prepared, AI students doing much better on the job than those that are poorly prepared? Mr. Duke, you indicated that AI is used in job skills. All right.
Dave Duke
Well, certainly knowledge work is one of the most impacted domains for AI-ready job skills. So I would say any knowledge work that you can think of, whether you're going to be an accountant or a consultant, students who can leverage AI to create work products the first day on the job are going to do better than students who do not. Students who are going into more hands-on jobs in the trades are less affected by that, but that's the domain that we see impacted the most.
In the knowledge space sector?
Right.
Dave Duke
It is. However, you're going to have a hard time competing against a student that already has those skills on day one.
James Moylan
Thank you, Mr. Chairman. Panel, thank you for your testimonies. And Dr. Du, first question for you. In your written testimony, you stated access to higher education in the United States has always been shaped by geography, economics, and the availability of support structures. I totally agree with you, and I don't think anyone knows this better than the people of Guam and the millions of Americans who call the territories home.
Small economies and geographic isolation often mean that territorial development lags behind CONUS. So how is McGraw Hill and other private entities working to expand access to higher education and critical workforce skills in geographically isolated parts of the US?
Dave Duke
Great question, and thank you for calling that out in my testimony. The first way that we're doing that is integrating AI-enabled product experiences into our products. And so if you buy our products, you are getting that access baked in. One of the things I wanted to talk about earlier, I wasn't able to get in, is that we keep bringing together the idea that an LLM is the same as AI-enabled product experiences embedded into an existing platform that a student might use. They're two different things.
They're managed in different ways. And so in terms of access and our ability to make those kinds of features, whether it is a features that help a student translate content into their native language, whether it is the ability for student to get extra support after a faculty member is not around late at night, we bake those into our products so that that access extends to anyone using our platforms. That's primarily how we do that.
James Moylan
Thank you, Dr. Du. Mr. Horn, you used electricity and factories analogy to describe using AI in the workplace and how innovation requires us to rethink traditional processes. But it's also true that AI should not come at the expense of human dignity. Pope Leo recently stated, work is not considered simply as a problem to be dealt with or a means of generating income, but a fundamental good for the person, a principle of economic activity and the key to the entire societal questions. Through work, human beings bring their freedom, creativity, and the capacity for cooperation into play, contributing to the culture and moral evaluation of society.
Thought and creativity should not be replaced with ChatGPT, nor should we see humans displaced because of mass adoption. Instead, AI should enhance learning outcomes, job performance, and see our country become more competitive rather than complacent. So how can institutions, industries, and government come together to work towards measurable progress to keep the US as the leader in education and innovation?
Michael Horn
Thank you for your question. I share the concern and what I would say is a few things. Number one, fundamentally in education we are about building human capital and the human dignity of each individual. And that's the center of the enterprise. And so everything should start from that point.
The second one I would say is there is a view of the future of work where it becomes fewer people working. I don't share that view. I think that what we have seen throughout history is that when technology makes something abundant, the things around it all of a sudden seem scarce and human ingenuity leans into that. And so much so the importance of making sure that individuals in education institutions have opportunities to do real work with real employers, build those connections, build that understanding of work, what work looks like, so that they can become the creators of those jobs and companies and processes in the future that will power the American economy. And I think the more we give those opportunities, the more we will continue to signal that the human is the center of that.
James Moylan
Thank you. One final question for you, sir, Mr. Horner, building off your point that colleges need to redesign their models as AI reshapes education and the workforce. For students in places like Guam, where geography can limit access to specialized programs and employer-connected learning opportunities, how can AI help overcome geographical barriers and prepare students for careers and opportunities of the future?
Michael Horn
It's really interesting. When online learning burst on the scene, it all of a sudden created much more access regardless of where you were to expertise. What I think AI is now doing is, for the first time, we've heard a lot about virtual reality simulations and things like that over the years. We can actually now create those at scale that bring it to people wherever they may be and give them access to those real embedded opportunities regardless of what the local ecosystem looks like.
Thank you, panel. Thank you, Mr. Chairman.
Speaker 40
Thank you. We'll now move toward move to our closing closing remarks. I'd like to recognize Dr. Adams for her closing remarks.
Alma Adams
Thank you, Mr. Chairman. And let me just, first of all, thank all of the witnesses again for testifying today. Today's discussion has reinforced a simple reality. AI is is not a future issue. It's already affecting how students learn, how faculty teach, how institutions operate, and how employers evaluate skills and qualifications.
The decisions we make now will will shape whether this technology expands opportunity or deepens existing inequities. No one is suggesting that we should ignore innovation or try to stop technological progress. The question is whether we will put in place the the safeguards necessary to ensure that that innovation serves the public interest. Students should not have to worry that an opaque algorithm will determine their educational opportunities. Faculty should not be left without guidance as as these tools become more integrated in classrooms.
And institutions should not be forced to navigate these challenges alone. As AI continues to evolve, guardrails become more important, not less. We need transparency, accountability, strong protections for student data, robust enforcement of civil rights laws, and clear guidance that helps institutions use these tools safely and responsibly. Higher education has always been about preparing people for the future.
I spent 40 years in the classroom at Bennett College in Greensboro, North Carolina. And I know that our responsibility is to ensure that the future we are preparing them for is one where technology strengthens learning, expands opportunities, and respects the rights and dignity of every student. That is the work before us, and it is the work that we cannot afford to delay. So I thank the witnesses again. I thank you, Mr. Chairman.
I yield back.
Burgess Owens
Thank you. I'm going to highlight what the chairman of the full committee said. This is a hearing I wish all of my colleagues could sit in and listen to. And I also can see the bipartisanship we have here. We all have concerns, make sure there's a balance between opportunities and also the dangers of this process.
But I came across a quote just a couple days ago: productivity, productive capacity is power. I'm so excited for teachers like my parents. My dad was a professor at Florida A&M for 40 years, mom was a high school teacher. And they, their eyes, they just were so excited by seeing the eyes of their kids that they're teaching that just open up, their confidence grow, and that was their mission. I'm so excited about teachers like that and the future we have as we produce really, I think, one of the most productive generations coming if we make sure we balance this right.
I wrote down just a couple words as we were going through this, I just wanted to share with you guys. Words like disruption, sharing of OpenAI. We even got into electricity, we talked about paradigm shift for electricity. Personalized learning at scale. Ubiquitous access, and we talk about colleges like the HBCUs that would need this.
Knowledge work, critical thinking, human dignity. These are all words that we can now say we have a great opportunity to embrace. I cannot wait to see the productivity we're going to have with our kids coming into the future as they begin to understand how to critical think and work in this process. And I want to thank you also, you and your industry. This is a remarkable time, and we need to have people that are smart, that are passionate, that have the intelligence, the wisdom to know the difference and also to know that you can lean in and have this collaborative effort to make sure that we all grow as a country.
This is a very important time for our country's future, and this process you're going through right now and what you represent will make sure we're leading in a good way. So I want to thank you for all the work you're doing, for work that we'll be seeing in the near future. And I'd like to again thank the witnesses for taking the time to testify before the subcommittee today. Without objection, and there being no further business, the subcommittee stands adjourned.