The unchecked proliferation of AI-altered messages has displeased some Democrats, potentially spurring moves on Beacon Hill to regulate how and when AI-generated content can be used in campaigns.
Massachusetts House leaders are introducing a pair of bills for a vote on Wednesday. It also includes a bill that would prohibit campaigns from disseminating synthetic media that is “substantially deceptive and intended to cause harm.” [a] Candidate reputation” Within 90 days after the election. The Republican-led proposal would require campaigns and PACs to disclose when AI will be introduced into campaign “communications.”
“People want assurance that what they’re consuming and what they’re reading is what they see,” said state Rep. Dan Hunt, chairman of the Election Law Committee.
“When you open social media, you pretty much expect it to be fake or AI. If you see someone throwing a football from the end zone outside the stadium, it’s fake,” the Dorchester Democrat said. “But when you use someone’s image or likeness, it’s not always easy to tell.”
Hunt said the House’s goal is to have legislation in place for elections this fall. But some Senate leaders have indicated they intend to further regulate how AI is deployed, including banning its use annually.
“If we don’t stop it, I think this is going to become part of the campaign and something out of the box,” said state Sen. Barry Feingold, co-chair of the Legislature’s Emerging Technologies Committee.
The Andover Democratic Party cited the variety of AI content targeting Healey as an example of why lawmakers should place limits on AI-generated images ahead of this fall’s election.
“Looking at what happened with Governor Healey, we want to act sooner,” he said.
Currently, about half of U.S. states regulate the use of AI in political messaging, with some Republican-led states (Texas) and Democratic-led states (Minnesota) outright banning AI as elections approach. But Massachusetts is not among them.
In 2024, the state Legislature passed the bill and Mr. Healy signed it. This aims to prevent so-called deepfakes in campaign videos and ads by prohibiting campaigns from spreading deceptive images with “actual malice”. In other words, knowing that something is false and publishing it.
However, this measure was designed to be only temporary. The law expired a year ago, and given the timing of its passage, it was never actually enacted in the state election.
One bill the House is scheduled to vote on this week includes similar language without an expiration date. It also includes an exclusion for material that “constitutes satire or parody.”
State Sen. Michael Moore (D-Milbury), who sponsored the 2024 bill passed by the Legislature, introduced a similar bill this session. He said the first iteration was temporary because some lawmakers were concerned about the language having “long-term effects.”
But he said the urgency to pass something is growing as campaign attacks increasingly move away from using authentic but unflattering photos of opponents. Publish images and sounds created in their entirety.
“Right now they’re treading water,” said state Sen. Michael Moore, D-Millbury, of Republicans’ attacks on Healey.
He gave these examples as follows: “That’s exactly why you should introduce something right now. They might say it’s a parody or a comedy, but the people watching it don’t necessarily know that.”
Brian Shortsleeve, one of three Republican candidates vying for the party’s gubernatorial nomination, released a video depicting Healy as a Grinch and a vampire. Federal Unity PAC, another group supporting Short Sleeves, posted what appears to be an AI-generated video on Christmas morning of a woman resembling Healy rushing to open presents, only to find coal.
Short Sleeve’s camp said the video was “obviously” a parody intended to get the point across and was not intentionally deceiving viewers. One focuses on a news story about how Mr. Healy’s former aide, Lamar Cook, received $31,000 in compensation after his arrest on cocaine trafficking charges. The video shows a moment where Cook stands with his hands behind his back in an actual image taken in court, before the AI-generated video begins to depict him. He dances with joy as money falls from the ceiling.
Short Sleeve’s campaign says its AI videos have been viewed a total of 350,000 times online.
“Based on the feedback we received, voters understood the joke and liked it,” said Patrick Nester, a spokesperson for the Short Sleeves campaign.
Nester said the campaign’s policy is to include a disclaimer: “Whenever AI is used to represent a real person in a way that is not obvious to a reasonable viewer.”
But most of Mr. Shortsleeve’s campaign clips do not include such revelations, including a satirical “radio ad” released in late January, the day Mr. Healey officially launched his re-election campaign. of The campaign includes a disclaimer in the video posted Friday, informing viewers of its features. An AI-generated replica of Healey’s voice listing Massachusetts businesses that have reported employee layoffs.
The content is a parody, the memo says, “but the information presented is factual.”
“Brian supports clear rules that respect the constitutional right to free speech, including the Supreme Court’s clear protections against the use of parody and satire, and our campaign is already following these principles,” Nestor said in a statement.
Nester said Swiftcurent created the campaign’s AI-generated content. Short Sleeve has paid the company more than $14,000 since the summer.
The Federal Unity PAC supporting Shortsleeve operates separately from his campaign. But the company is also leaning heavily into AI-generated content, typically targeting Healy, for which it has already spent heavily.
The super PAC reported spending nearly $430,000 through the end of December, including $34,000 in production costs and digital ads opposing Healey. In October, it released digital ads and social media posts that featured images of the Democratic incumbent wearing a sombrero and criticized him for offering “free hotels to immigrants,” among other things.
Republican operative and super PAC chair Lydia Goldblatt did not respond to multiple requests for comment.
A spokesperson for Mr. Healy’s campaign declined to comment on whether she supports putting limits on videos targeting her or the AI content that can be used in campaigns.
The proliferation of AI is happening far beyond Massachusetts politics. President Trump’s media team has long been pumping out AI images. Critics say some images depict cartoons of the president himself, while others, such as an edited photo of a civil rights lawyer, distort reality in dangerous ways.
California Governor Gavin Newsom is also increasingly using AI, primarily as a way to mock Trump. New York Governor Kathy Hochul, also a Democrat, has called for a ban on campaigns disseminating AI-generated images of people ahead of elections.
So far, 26 states have passed some sort of law targeting the use of “deepfakes” or “deceptive” media. The National Conference of State Legislatures, a nonpartisan organization that tracks bills across the country, says most of them in politics require disclosure at the time of use.
Some states, including Rhode Island and New Hampshire, passed laws after robocalls in New Hampshire imitated President Biden’s voice and urged voters to skip the state’s presidential primary in January 2024. The political consultant who sent the robocall was ultimately acquitted of charges of voter suppression and impersonating a candidate.
The spread of AI through campaign messages is worrying those who study AI.
“If people can’t tell what’s real and what’s fake, we should filter out the fakes to establish election credibility,” said Jim Glass, a senior researcher at MIT’s Computer Science and Artificial Intelligence Laboratory.
Matt Stout can be reached at matt.stout@globe.com. follow him @mattpstout.
