Klobuchar announced that the Rules Committee will mark up her bills to address AI in our elections.
WASHINGTON - At a Senate Judiciary Subcommittee hearing titled “Oversight of AI: Election Deepfakes,” U.S. Senator Amy Klobuchar (D-MN) urged action to address the impact of artificial intelligence (AI)-generated deepfakes on elections.
“Like any emerging technology, AI has great opportunities but also significant risks,” said Klobuchar. “This is the one right before us [...] and we have to put rules in place.”
A rough transcript of Klobuchar’s remarks and questions is available below. Video is available HERE for download.
Senator Klobuchar: Thank you. Well, thank you, Chairman, and thank you, Ranking Member Hawley, for this important hearing, [and] this opportunity to keep this on the front burner.
As Senator Hawley just said, we cannot wait. We are scheduling a markup of our bill, and we are going to have to work… it's the only committee that both leaders are on… fun committee to chair, and so I will seek Senator Hawley's help and others on our bill, which includes Senator Coons and Collins, Senators Bennet and Ricketts, and a whole lot more support on both sides of the aisle to get the votes. Not just to, you know, we can obviously pass it, but I'd like to get a really strong vote coming out of committee so we can immediately get this thing heard, because we really can't wait. The elections are upon us.
And like any emerging technology, AI has great opportunities but also significant risks. And this is the one right before us, as well as other issues related to scams, and we have to put rules in place, and we can't let the same thing happen as every one of the four of us has been out front on this. As happened with Section 230, and what happened when they just acted like these companies were things in a garage, and now they're humongous monopolies, and now we are all challenged and trying to get these bills forward. Whether it's on fentanyl, whether it is on child pornography, whether it is on competition policy. We have to move these.
The fake robocall. I hadn't actually heard it myself, so thank you for that. And it is just impossible to tell that that's not Joe Biden as it was impossible to tell one video that ran during the Republican primary that wasn't accurate involving Donald Trump. That also wasn't accurate. Or we had an Elizabeth Warren video in which she says “Republicans shouldn't vote,” that wasn't her, but you couldn't tell. We had in Minnesota, and this is not AI, but it just shows how devastating this can be. We had a photo the day after the heroes, the two police officers in Burnsville, Minnesota, were killed after rescuing seven kids, and then the paramedic was killed who was performing CPR. A photo of an actual rally picture from 2022 that I was kind of in the background on started going around. At the same time, there's some kind of Russian photo going around saying that I fund Nazis in Ukraine. That's been going around for three years. This photo had a red circle around me in the background, and then they put “defund the police" signs in the hands of the people at the rally that were never there. So they were literally using, the people who did this, I personally think it was foreign interests, took a photo and put those “defund the police” signs, after these officers have been killed. And to their credit, X and Meta, put “altered content” with the big sign, but it took us about, you know, a day to get all this down. It was going all around the internet. That is actually not AI. That's a real photo that they doctored, and people thought it was real. It looked real. And so this kind of thing is just going to keep happening and keep happening unless we take immediate, immediate action.
Eleven states, including my own, have enacted laws to address these threats to our elections, and that's great, but it doesn't cover federal. Some of these states are–they're not all blue states, purple states, people are taking action as seen by the bipartisan nature of our legislation. And we also need disclaimers on other ads that aren't deepfakes, and that's a bill that we will also be marking up in the Rules Committee. So I want to thank my colleagues for doing this. I want to thank them for their willingness to stand up on this issue, and [I] look forward to hearing the testimony of the witnesses. Thank you.
…
Senator Klobuchar: Right now, I'm going to focus on elections, but I will say those were startling numbers, Mr. Colman, and I think it is just what we're seeing both real people and AI-created people, it's one of the reasons that we got the SHIELD Act through here, which is not the liability issue that I also support, that Senator Hawley was mentioning, but also getting the information to law enforcement and the like, to be able to make it easier to go after these perpetrators. And we can just sit here and do nothing, we can pass resolutions, but unless we empower people to go after these cases and then equally make liability, it's just gonna get worse and worse, and at some point, the public will have had it. I don't know if that's what–this year–but it's going to happen. And so, I keep telling my colleagues this. So let's go to a few things here, the bill that I mentioned that Senator Hawley and Coons and Collins and Bennet, Ricketts, others [and] I have. Could you tell me, Mr. Colman, how AI has this potential to turbocharge election-related disinformation and why we can't just rely on the disclaimers and watermarks? I think you can do that for a set of it. I don't think you should do it for all uses of AI, and we have a labeling bill that I think differentiates that, but for this really bad stuff that Secretary of State Scanlon was referring to, tell us why it's not enough. This is softball but to run the whole thing and have a little label underneath when they think it's the actual candidate, but it's not.
Ben Colman: We agree on that. I think that to paint the larger picture, what we saw during the primaries was a single static deepfake, pre-recorded, kind of a one-to-many attack. It didn't change. It wasn't even live. Imagine a world where that was a one-to-one attack where instead of it being pre-recorded, it was actually live, and instead of being from one-to-many, it was one-to-one where it's coming from your husband, your wife, your boss, saying “Hey, Ben, we need you in the office of 6 am, I know it's a voting day” or to an election official “Hey, we're moving your precinct, we need you to be across town three hours away.” And that's where this is gonna go. It's not going to be a single pre-recorded, you know, arguably medium-level deepfake. It's going to be a real-time custom deepfake in conversational language. Have people do all kinds of things at all levels of the election system.
So, on our side, we see this as a massive issue, not just in the U.S. but globally, and what's great here is on the dais we have different technologies, all solving very much the same issue. It's all possible now. We have large companies. We work with large banks, large government groups, large media organizations that are thinking long-term and already solving for this. We have banks scanning incoming phone calls, every single phone call. We don't have anyone protecting average consumers, my parents, my grandparents, they just don't stand a chance. With other technologies, whether it's CSAM or, for example, a computer virus, they don't have to be experts. They don't have to tell ransomware or an APT, they just know that their email provider will actually block it for them.
We're looking for the bare minimum there, which is just letting us know that maybe something might be fake and then allowing us to decide maybe we don't want to see it, maybe it won't go viral. But right now, the things that are most extreme go the most viral, and platforms that do think about this are already solving that using technology like ours.
Klobuchar: Right. Very good. And I do want to note that drafting this bill, the deepfake bill wasn't easy. We had to look at allowing satire, right? And all these kinds of things within the framework of the Constitution, and having Democratic and Republican lawyers look at this, to figure out what gave us a chance. I just think if the platform can point to something, as opposed to laws that aren't quite on point, which eleven states have done for states, but not for federal and say, “we've got to take this down.” We're going to be in a much better place than we are with a little label that they may not even notice. Also the labels, you know, I think it's important for some of this, but I don't think it can be the only answer. Mr. Scanlan, you know, birthplace of democracy, no kidding, I spent a little time in your state there. I know you cherish democracy very much. Could you talk about what other federal support will be helpful in taking this on in addition to stronger laws?
David Scanlan: Secretaries of State, for a good decade now, have been dealing with misinformation and disinformation generally, and that takes on many different forms. And there's no question that today, voters receive their news in different formats than they did twenty years ago, and a lot of that news is electronic. It's on your cell phone. Many voters believe exactly what they see on that tech, on that format, on that media without question. So, in addition to whatever might be appropriate to help states recognize and put brakes on malicious technology in terms of deepfakes, I think we have to spend a real strong effort on the fundamentals of transparency, and helping voters, and educating voters on the election systems, and how they run, and what the checks and balances are that are protecting them in the polling place.
Klobuchar: I've always found it interesting, like those Baltic states on the border with Russia, they were putting up misinformation, lying, things, and they, over time, because of education, they kind of have seen through some of it. It is possible, but it can’t be our only answer because of what everyone's being exposed to, but I think it's a good point, and we have the election assistant, of course, commission, but I did want to say appreciate as a Republican Secretary of State how seriously you and the Attorney General and others in New Hampshire took this egregious breach with the guy that did an interview afterward. Maybe they should have hired a mime instead of a magician. But in the end, I just think that we've got to make clear there's consequences when this happens as well.
I have other questions. I don't want to go over my colleagues' time, I already have–I’ll ask you on the record, Mr. Ahmed and Mr. Gupta, thank you so much for being witnesses today. But I just..we have to be as sophisticated as the people that are messing around with our democracy and our laws, and that's why we’ve got to get these bills done. Thank you.
Klobuchar, Chairwoman of the Senate Committee on Rules and Administration with oversight over federal elections, has led efforts to address the threat of misleading AI-generated content and to increase transparency in our elections.
In September 2023, Klobuchar and Senators Josh Hawley (R-MO), Ranking Member of the Judiciary Committee’s Subcommittee on Privacy, Technology, and the Law; Chris Coons (D-DE), Chair of the Judiciary Committee’s Subcommittee on Intellectual Property; and Collins, introduced the Protect Elections from Deceptive AI Act, bipartisan legislation to ban the use of AI to generate materially deceptive content falsely depicting federal candidates to influence federal elections. This legislation has also been cosponsored by Senators Michael Bennet (D-CO) and Pete Ricketts (R-NE). Klobuchar also held a Senate Rules Committee hearing in September 2023 titled “AI and the Future of Elections,” highlighting the need to address the risks posed by AI to our democracy.
In March, Klobuchar and Senator Susan Collins (R-ME), Vice Chair of the Senate Appropriations Committee, introduced the bipartisan Preparing Election Administrators for AI Act to require the Election Assistance Commission (EAC), in consultation with the National Institute of Standards and Technology, to develop voluntary guidelines for election offices. These guidelines will address the use and risks of artificial intelligence (AI) in election administration, cybersecurity, information sharing about elections, and the spread of election-related disinformation.
In March, Klobuchar and Senator Lisa Murkowski (R-AK) introduced the bipartisan AI Transparency in Elections Act to require disclaimers on political ads with images, audio, or video that are substantially generated by artificial intelligence (AI). The bill also requires the Federal Election Commission to address violations of the legislation quickly.
In February, the EAC voted unanimously to assist state and local election officials in combating the spread of AI-generated disinformation about our elections by allowing election officials to use federal election funds to counter disinformation in our elections caused by artificial intelligence. This decision came after Klobuchar and Collins’ letter calling on the EAC to take action to address AI-generated disinformation in elections following the fake AI-generated robocalls using the president’s voice to tell people not to vote in New Hampshire.
###