State legislators tighten AI rules to combat misleading election ads
When artificial intelligence experts recently showed a gathering of state legislators a deep fake image, created by AI in early 2022, of former President Donald J. When Trump and Barack Obama were shown playing one-on-one basketball, the crowd laughed at how rudimentary it was.
The panel then brought out a fake video that was made just a year later, and lawmakers were shocked to see how realistic it looked.
Concerned by the increasing sophistication of false or highly misleading political ads generated by artificial intelligence, state lawmakers are struggling to draft bills to regulate them.
As primary voters head to the polls for the first time in 2024, the issue has become even more pressing for legislators in dozens of states who are returning to work this month.
“States know there have to be some regulatory guardrails,” said Tim Story, president and chief executive of the National Conference of State Legislatures. AI panel convened at a conference In December. “It's almost like trying to figure out what's happening in real time.”
The overarching goal, the lawmakers said, is to prevent what has already happened elsewhere, particularly in some elections abroad. in slovakia, deepfake voice recordingAs the leader of a pro-West political party falsely claiming to have bought votes, that party may have suffered a minor loss from a pro-Kremlin party. And last year, Florida Governor Ron DeSantis appointed former President Donald J. Released fake AI images of Trump hugging Dr. Anthony Fauci.
According to Public Citizen, an advocacy group, as of early 2023, only California and Texas had enacted laws related to the regulation of artificial intelligence in campaign advertising. keep track of bills, Since then, Washington, Minnesota, and Michigan have passed legislation with strong bipartisan support mandating that fact be disclosed in any advertising created with the use of artificial intelligence.
By the first week of January, 11 more states had introduced similar legislation – including seven since December – and at least two more were expected soon. Penalties vary; Some states impose fines on offenders, while others make the first offense a misdemeanor and subsequent offenses a felony.
state representative Julie OlthoffA Republican from northwestern Indiana attending the legislators' conference in Austin, Texas, said his background as a marketing and advertising business owner made him realize the potential dangers of people trying to manipulate images and words. Got it done.
his billFiled on January 3, any “fabricated media” using AI would be required to come with a disclaimer stating, “The media portraying the candidate has been altered or artificially generated ” The bill would also allow candidates who were the targets of AI ads to bring civil actions.
“People don't know how much to trust a source anymore, so I think that will help,” he said.
Several AI bills have been introduced in Congress, including One Led by Senators Amy Klobuchar of Minnesota, a Democrat, and Josh Hawley of Missouri, a Republican. But those bills would apply to federal elections, not state or local elections Robert WeissmanPresident of Public Citizen, which has filed a petition with the Federal Election Commission Take additional action against deepfakes,
“It's one thing to refute a lie or a mischaracterization, but to refute a concrete video or recording of you saying something, what do you do?” He said. “That's why we're seeing this widespread interest.”
Some legislators have considered banning misleading AI ads altogether. But political ads are generally given a lot of leeway about what they can say, and to avoid any First Amendment challenges, most lawmakers have focused on what the people who create the ads do. Disclose the information they create, create or transmit in legible text or clear audio. The misleading ads were created by artificial intelligence.
Many bills only apply to ads that are released 90 days before an election, when voters are paying the most attention.
minnesotaThe new law, enacted in May, targets people who use deepfakes to create sexual material without consent, or to harm a political candidate or influence an election, the state representative said. jack stephensonA Democrat who represents the northern Minneapolis suburbs.
A voice claiming to be Mr. Biden said, “No more mess.” “As my father used to say, 'Joey, you can't believe everything you hear.' not a joke.”
But it wasn't real. Instead, a friend of Ms. Tsernoglou, who had no experience with technology, used an AI voice generator, she said in an interview.
“He said it took him five minutes,” he said.
Ilana Beller, field manager for the Public Citizen Democracy Campaign, said the proposals have faced minimal opposition to date. Technology companies have also been generally supportive, while trying to ensure that they are not liable for unknowingly circulating unlabeled deepfakes on their platforms.
Among the half-dozen states that have introduced AI bills since December, kentucky It stands out because even first-time violators will be charged with a felony and could be sentenced to up to five years in prison.
One of the sponsors of the bill, State Representative John HodgsonA Republican from suburban Louisville said he thinks a fine of several hundred or thousand dollars would not be enough of a deterrent.
Watching as he took his pet sheep Sassy and Bossy to a live nativity scene during the Christmas holidays, Mr. Hodgson, a retired UPS Airlines executive, thought: “Imagine it's three days before the election, and “Someone says I've been caught having sex with a sheep and it's been delivered to a million voters. You can't recover from that.”