Chuck Schumer Talks About the Importance of Understanding AI

Good start.

Yeah. Thank you for coming. We just had a productive lunch and as you may have heard today, I gave a speech about one of the most pressing and consequential issues and innovations of our time, and that is AI.

(00:19)
I laid out my ideas about a comprehensive framework Congress can use to supercharge AI innovation in a safe and responsible way. I call it the Safe Innovation Framework for AI. I call it that because innovation must be our North Star. AI could be our most spectacular innovation yet, a force that could ignite a new era of technological advancement, scientific discovery, industrial might.

(00:47)
So, we must come up with a plan that encourages, not stifles innovation. But if people don’t think innovation can be done safely, it will stifle AI’s development and even prevent us from moving forward. So, that’s why my Safe Innovation Framework calls for certain guidelines, guardrails rather, for security, accountability, protecting our foundations, and lastly, explainability, which is one of the most important and most difficult technical issues in all of AI.

(01:24)
Congress has never had to deal with issues like AI before. Why is it different? Well, first, we don’t have much of a history with it. It’s not like labor or healthcare or defense where we’ve had a whole history that we can work off of. Experts are not even sure… Experts… Yikes. It’s quite a day. Experts… I had a good joke, but I [inaudible 00:01:57]. Can I tell my staff and get it off my chest? No, okay.

(02:04)
Congress, as I said, we don’t have the history and the experts are not even sure which questions policymakers should be asking. In many ways, we’re just starting from scratch. So, because AI is so fast moving, so complex and so outside Congress’s expertise, I also announced in addition to our SAFe framework… And you can get a copy of the speech, we elaborated on what all of these S-A-F-E and innovation mean.

(02:34)
But I’m going to invite the top AI experts to come to Congress and convene a series of, first ever, AI insight forums for a new and unique approach to developing AI legislation. These insight forms are the first of their kind. They have to be, because AI moves quickly, will change our world dramatically, is deeper in its complexity and lacks legislative history.

(02:59)
Our job as legislators will be listening to these experts on nine different issue areas within AI. We’re going to call the leading experts and have them sit and discuss with each other, how do we deal with these issues? Let me give you one quick example.

(03:19)
Company driven AI or open AI. Now, if it’s company driven, how do newer startups, newer groups get that information? On the other hand, if it’s open, can our enemies get hold of it? Can’t rogue actors get hold of it? Can’t groups that want to use it for financial extortion or ideological purity get hold of it?

(03:47)
So, these forums will have the experts come together and do in a few days what might take months or years to come together if you went to each of them separately and they all just said what they thought. So, we need an all-of-above approach, but our committees have to continue to be the key drivers. These forums are not to replace or supplant the committees, they’re to give the committees the fuel they need to actually draft good, strong legislation.

(04:18)
And we need bipartisanship as we work in all of this. That’s why I established a group of senators, Senators Heinrich, Young and Rounds, and I thank them for their work, and as well as relying on committee chairs. And last week at our chairs meeting, I asked every one of our chairs to reach out to their ranking member and see what they thought they could do in their committees on AI. Okay.

(04:42)
Now, no question, this is ambitious. And I have a word that I use to describe what we’re doing, humility. We need humility as we proceed because this is hard. We’re going to come work, but we have to do it or the alternative is worse. We’re going to work very hard to come up with comprehensive legislation. Because it’s so important, we’re going to do everything we can to succeed. Success is not guaranteed. AI is unlike anything we’ve dealt with before and it’s difficult. But even if we can find some solutions and create a degree of consensus to deal with some of AI’s many challenges, we got to pursue it. So, that’s what we’re doing there.

(05:27)
Now, second issue. Today, the Senate will reach a major milestone. In a few minutes we’ll confirm our hundredth district judge under President Biden with Natasha Merle for the Eastern District of New York. Many of these hundred judges have knocked down longstanding barriers in the halls of justice. The first Muslim district judge. The first woman of color to be district judge is in Maryland, in Oregon. The first openly LGBTQ judge in Puerto Rico, and the list goes on and on and on. And you know the statistics, how much diversity we’ve sought, not only demographic diversity, more than two thirds women and people of color, but also professional diversity. They’re not just lawyers from law firms or prosecutors, it’s much more across the board.

(06:16)
So, hitting the milestone of a hundred judges are significant. Our bench is now far more balanced and diverse and more experienced than the one we had a few years ago, and we’re going to keep at it. Okay, questions? Yes?

Thank you. [inaudible 00:06:30]. What are the worst case scenarios associated with AI challenges?

There are lots of different… There’s so many different issues that face us. How do you deal with… How do you… The two big questions are, how do you encourage innovation and yet have guardrails that deal, that prevent AI from going off track? And the guardrails are on these categories, how do we preserve democracy? How do we keep our elections fair? AI would allow somebody to come up with a very convincing false image of any Senate or a Congressman, Democrat or Republican, and they send it to a hundred million homes. You can’t put that genie back where they say ridiculous things.

(07:14)
Elizabeth Warren brought up at our lunch today that there’s an AI video of her saying Republicans shouldn’t vote, but looks like her, sounds like her, acts like her. So, we have so many different areas. And the toughest, and this is a little bit of a lesson, is what I call, explainability. When AI… When the AI’s process spits back a paragraph, a sentence, an answer, you want to know where it comes from and you want to know why it was chosen and something else wasn’t chosen. I’ll give you a facile example.

(07:47)
So, there’s this great diet and it’s very convincing and written to convince tens of millions of people, but it says, “Eat a lot of hot dogs.” Well, if you knew it came from Oscar Mayer, or you knew they took the words from a Oscar Mayer ad, you’d know something. It’s one of many examples. Yes?

What is your reaction to the Hunter Biden DOJ detail?

Look, I think that there should be no political or outside interference and the law ought to go forward. Yes?

Mr. Schumer, do you agree with President Biden’s assessment that President Xi is a dictator? And do you think that was a responsible thing for President Biden to say?

Well, look, I’d say he has many, many autocratic tendencies. They are not an open and full-fledged democracy to say the least. Yes?

[inaudible 00:08:38], an unusual question. My question has already been asked. You are [inaudible 00:08:41] chairman, do you believe that the press won’t be able to have devices-

The what?

Should we be able to have devices in the senate gallery [inaudible 00:08:50]-

I’d have to defer to Senator Klobuchar on that. It’s a while back since I was rules chairman. Yes?

Were any restrictions [inaudible 00:09:01]-

I can’t hear you.

Were any restrictions on [inaudible 00:09:04] access [inaudible 00:09:05] here [inaudible 00:09:08]?

Look, the committee is meeting now and we’re going to try to do everything we can to get a good NDA bill passed. Obviously I’m opposed to abortion restrictions in that Bill. Yes? Bolton and then-

Judiciary’s new Chairman, Dick Durbin and Sheldon Whitehouse have been marking up Supreme Court ethics legislation after July 4th recess. Would you bring that to the floor or would you have to see at least 10 republican votes for it before-

Well, look, I support their Bill. It’s going to come out of judiciary. Let’s see what happens there first. Yes? Last one.

Is Julie Su going to get a vote in this chamber?

Look, she is a very good nominee and we are working hard to try and get her confirmed. Thank you, everybody.