News from Vermont Public
Vermont's secretary of state is raising concerns over the use of AI in politics
Artificial intelligence is becoming more and more ubiquitous in our daily lives, including in the world of politics. AI is creating new and highly targeted ways to reach specific voters and campaigns see those tools as a cost-effective way to communicate with very narrow demographic groups. At the same time, some are concerned about how AI could be used to undermine voters' trust in our systems of government. That includes Sarah Copeland-Hansis, Vermont's Secretary of State. Vermont Public's senior political correspondent Bob Kinzel interviewed her recently, and he joins me now to share what they talked about. Hey, Bob. Hey, Jen. Bob, what are some of the concerns that the Secretary of State has raised about the use of AI in politics? Well, Jen, I think she recognizes there are both some possible benefits of AI and some serious concerns, as you mentioned. Now, let's look at the plus side. It can be a valuable tool for campaigns to target their message to voters. They can identify people based on available data and then tailor their message. For instance, if someone is interested in climate change or reproductive rights or immigration issues, there are an endless number of messages that can be sent out at a very low cost. Now, at the same time, the Secretary of State is sounding an alarm that these AI tools have to be watched very closely, and that's because she wants people to know they can be used in a deceptive and misleading way. And I think her basic concern is the ability of voters to tell the difference between fiction and reality. Now, think about it. A fake video from a real one. A fake photograph from a real one. The technology is so good that it can be almost impossible for someone to tell the difference between the two. And I think we've all seen and heard stories in the news about, you know, some candidate being shown to have said or done something completely outrageous, only to find out that, you know, that was an AI fake. Jen, this is also a worldwide issue where some governments, for instance, China and Russia have been accused of generating misinformation to try to influence the outcome of elections in other countries, including the presidential election in the United States this year. And Bob, am I correct that we are still living in a world in which these videos or the images do not indicate that they are AI generated? Not yet. That's exactly true. And that's going to be a big issue coming up with legislation in both the federal level and at the state level. Okay, Bob. So we've been hearing about fake videos and fake photos. Of course, AI can be used to spread misinformation using audio. And don't we have a recent example of that? We certainly do, Jen. It's the so-called Biden-Robo call. It came right before the New Hampshire Democratic presidential election. It's probably the most famous. Now, this was an election where there was a dispute between New Hampshire officials and the Democratic National Committee about the scheduling of the primary. So President Biden did not register as a candidate for the ballot. Instead, his supporters were encouraged to write in his name. The robo call was allegedly used to try to derail the Biden write-in effort by discouraging people for voting for him at all. And the message was delivered by Biden himself. So imagine being a voter in New Hampshire and getting this call. Voting this Tuesday only enables the Republicans in their quest to elect Donald Trump again. Your vote makes a difference in November, not this Tuesday. If you would like to be removed from future calls, please press two now. It's estimated that the call went out to as many as 25,000 people and the perpetrator initially faced a fine of $6 million. Now it's currently under negotiation. Now the person responsible for the call claims he did it to raise public awareness of the dangers of AI. So the courts will now have to resolve this issue, but it sure did get people talking. It certainly did. And we will keep following that story. I'm curious, Bob, what advice does Secretary Copeland hands us have for folks if they come across something they think might be an example of AI political disinformation? Jen, her basic message is trust your instincts. It's very similar to the age old consumer protection advice that if it's too good to be true, it probably is. But the most important really is that gut check to say, you know, something doesn't look or seem or sound quite right. I'm going to look a little bit closer and see if I can tell whether this is real. So Jen, if you get a robo call where you hear Senator Bernie Sanders and a very distinctive voice telling you that he's had a major change of heart and is now enthusiastically endorsing Republican presidential candidate Donald Trump, that should set off some warning bells. It certainly would. Bob, I mean, what steps if any are being considered to address these concerns at the legislative level? Well, there are some, but there are really no easy answers. Copeland Henson says she wants to work with the Vermont legislature on this issue and focus initially on disclosure efforts. Certainly to be requiring at a very minimum that if you're using an AI generated image in campaign material that you disclose that that is AI generated. But there are probably stronger protections that are being developed out there as more and more states tackle this emerging issue. Jen, the Federal Elections Commission is also looking at this issue, but there are questions within the FEC if it even has the jurisdiction to regulate these actions. And there are also people who believe that the regulatory cure to this issue is worse than the problem itself because there's some serious First Amendment issues here. So if any regulations are put into place in the coming years and we won't see them for this election, the final outcome will likely have to be resolved by the courts. Bob Kinzel, thank you so much for highlighting this important issue and we'll continue to stay tuned to your reporting. My pleasure, Jen.