Speeches

Rep. French Hill (AR-02) spoke at a Ripon Society breakfast about Artificial Intelligence (AI) and his leadership in Congress on AI as Co-Chair of the House Financial Services Committee's AI Working Group and as a member of Speaker Johnson's AI Task Force in the House. 

To watch Rep. Hill's full remarks, please visit HERE. 

Rep. Hill's remarks as delivered:

Well, it shouldn’t be a big surprise to anybody to come to Capitol Hill and talk about artificial intelligence. The interns are trying to figure that out. You can go to ChatGPT and ask the average IQ of a congressman, and you’ll see how that turns out. But this is an important topic because, as Chris so well laid out: it’s about financial services. It’s about the future of financial services, and Jim, in his introduction, talked about TR’s look to the future. You can’t ever go backward. You’ve always got to move forward. You can’t change events. It’s about how you manage your own behavior in events, and human progress is ever-changing, ever-moving, never static, and AI is one of those things.

When these moments happen, you definitely have people reacting in challenging ways. Fear is a typical one. People freeze up and don’t do anything or they try to put a rock over the top of it and not let it advance or move, and that’s not realistic. If you go back to the 1990s, if that’s what former Congressman and former SEC Chairman Chris Cox had done back in 1996, if that had been the direction of the Congress in the '90s on telecommunications policy, you wouldn’t have all that computing power in your pocket. You wouldn’t have been able to see the cheap, affordable, in-your-pocket phone experience that you’ve had since the ‘90s.

Alltel Information Systems and Alltel corporation, headquartered in Little Rock and now part of Verizon, were one of those pioneering public companies right in the Second Congressional District that created the competitive market for cellphone technology. You wouldn’t have had that. You would not have had the free and open Internet that allowed you to have a dial-up AOL account, an email, and begin surfing the net and determining just how great of a future it is. You wouldn’t have had the chance to either own Pets.com or Amazon in 1998.

You wouldn’t have seen global commerce blossom if Congress had decided in 1996, “No, this is insane. You know there is pornography out on this thing called the World Wide Web. I don’t know how it works. How do we protect our kids from it? How do we protect people from sending you an email that’s from a Nigerian prince? We need to regulate it. We’ve got to stop it. We need to put a rock on top of it.” Did we do that? No. Have you seen the entire globe prosper and benefit from a free and open internet since the ‘90s and lower, cheaper, and more effective computing power and cellphone services in your pocket? Yeah, I think so.

So, we are now at that point when you think about that. Whether you are talking about blockchain, in the other hat that I have in FinTech, on how we’re going to write apps in the next level of the internet on a blockchain. The same thing is true about artificial intelligence. So, we need to get this right, and I want to say on behalf of the Financial Services Committee that nobody has worked harder on understanding this for five years than the members of our committee.

You go back to 2019, Maxine Waters, then Chair Waters, worked with the committee to set up a FinTech Task Force and an Artificial Intelligence Task Force. I was pleased to chair both of those on behalf of the Republicans. Steve Lynch chaired the FinTech Task Force, and Bill Foster chaired the AI Task Force going back to 2019. Now, Foster and I serve on Hakeem Jeffries’s and Mike Johnson’s Artificial Intelligence Task Force for the whole House.

Steve Lynch and I co-chair the House Financial Services Committee working group on AI, and we’ve just put out our report this week. You can find it on the House Financial Servies Committee website, and it’s important that we in this country get this right because of our innovative spirit, our global position to help lead in technology, and our commercial private sector. There’s a growing effort to do that.

You’ve seen Chuck Schumer appoint an AI Task Force in the Senate. We have the one in the House. The White House has issued an executive order on this subject, and countries around the world are all grappling with what to do about artificial intelligence, including our friends in Europe who are intent on regulating it and setting up a lot of rules around it. The goal we had for our working group was to first ask the regulators and supervisors in the agencies of our jurisdiction, Financial Services’ jurisdiction, and HUD, “How are you using it in your agency?”

Some agencies had crickets in that conservation and others had engagement. Then, we asked the regulators, supervisors, and HUD, “How are you looking at this for the private sector entities and players that are in your domain? How are you setting up your guiding principles as a supervisor, a regulator, or a cabinet agency to oversee AI?” That was our very first working group of six that we did in the committee, it was pretty informative, and everybody, Steve Lynch, the Democrats, and the Republicans all listened to that.

We had HUD, SEC, CFPB, and the bank regulators. They were all principles-based, and I think that’s precisely the takeaway in thinking about AI oversight. Whether it’s in your business, your corporation, the Department of Defense, or as we think about oversight in our role in Congress, it has to be principles-based, meaning they weren’t going to stop you from experimenting with AI. They, the regulators, were not going to say to Bank A, “You can’t use AI in your business.” We’re not going to say that, and I think that is a decent starting point for this whole discussion because AI offers immense productivity gains, and we want more output per worker, which is what productivity is. AI offers that potential at scale.

We want people who are more compliant with the rules and regulations offered to protect consumers, businesses, and fair play between commercial businesses. AI offers the ability to do that more effectively than a group of watchers who are human and stare at computer screens all day. It’s a tool that makes the human leader, the human compliance officer, the human lender, and the human consumer advocate more effective. It is not a replacement for that person. It makes that person more effective at scale, faster to do more complex tasks, and leaves those decision-making tasks to the boss, to the person, to the human.

This technology is just like the game changer it was for me in high school chemistry when we didn’t have to use the slide rule anymore and we could use an HP12C or whatever the heck it was called. Jim remembers this because he gave it up for these stone tablets that he would use.

I view this as a tool to be used by American, Japanese, and global businesspeople, scientists, government leaders, and compliance officers to make their lives better and more effective. Now, I’ve connected the dots between being principles-based and all the concerns people have about unleased, generative AI.  I’ve linked those because I’ve kept referring to the boss: the human.

The second guiding principle: “Is there a human in the loop?” There is a person in the loop because, in the financial services space and for insurance capital markets activities, commercial banking regulations revolve around all aspects of real estate. From buying and selling a home, to financing it, to selling that mortgage in the secondary market you’ve got a lot of rules… a lot of rules. They aren’t really disrupted. They are to be followed. So, they are keeping a person in the loop. That comes to the third key point that we took away from the work we did on a bipartisan basis in the working group: because you use AI, you’re under no exemption from all the laws that you’re supposed to be complying with.

If you are using a piece of parchment and a quill pen, you’re under an obligation to comply with fair lending, fair credit reporting, state insurance regulations, and housing protections. The same is true if you use an AI machine learning algorithm to advance your work. I think that is the right way to think about this.

AI is a tool to be used to boost productivity in all kinds of extraordinary scientific, business, commercial, and personal interactions. It should be principles-based. We should let this be experimented with and used to make our lives more effective, whether it’s on our personal phones with ChatGPT or determining a new compound that will cure ovarian cancer.

One of the most inspirational aspects of this work has been that I’ve gone to MIT twice. Once with the House Intelligence Committee, where we go up and retreat at MIT and Lincoln Labs on things of national security related to quantum computing, artificial intelligence, the greatest advances we have in science, and how those impact the national security space.

Steve Lynch is a great representative of greater Boston, and I invited him up there for financial services to walk through the research at MIT on AI’s use in financial services. I walked away with amazement at the capability we have to map the human genome. Now, with machine learning, we can go through every conceivable compound and mix of proteins and determine what a drug would look like. That would take humans years to go through those permutations without using algorithmic machine learning. This data analytics was made a reality by artificial intelligence.

We will have new drugs and all kinds of things that we would never have if we didn’t have the power of large language models and machine learning which are using data sets designed by scientists at a major university like MIT. The other thing I walked away with is the humility of this.

This is not ready for prime time. There are many errors. The adoption model for AI use has skyrocketed. Only since 2017, when an AI image model on a university campus at Stanford, for example, couldn’t even figure out if that’s a car in the picture or not. It’s machine learning, so it is generative in the good sense that it is constantly trying to improve itself.

But it is learning, so it is not perfect, which is why you want the human in the loop. That’s been reiterated to me. In my experience, when talking to the private sector, people are being very cautious about this. They are very cognizant. They want a human in the loop. They’re very cognizant that they have all those laws and regulations to comply with and that they are still responsible for that, not some anonymous algorithmic solution to a question someone put into a chat box.

I hope you all enjoyed our report. Speaker Johnson and Minority Leader Jefferies will have a report this fall that outlines our viewpoints of the Congress at large, a much broader context than financial services, but Dr. Foster and I certainly enjoyed that experience and how we have helped bring the financial services point of view to that group as well.  I think I’ll stop there Jim.