As deepfake videos and audio permeate political life, California enacted its toughest regulations to date in September. The law would ban political ads that contain deceptive, digitally generated or altered content unless the ads are labeled as “manipulated” within 120 days of an election. ”

On Wednesday, a federal judge temporarily blocked the law, saying it violated the First Amendment.

California still has other laws in place to crack down on deceptive campaign advertising, including one that requires candidates and political action committees to: disclose If the ad uses artificial intelligence to create or significantly modify the content. However, the preliminary injunction granted against Assembly Bill 2839 This means there is no broad prohibition on individuals using artificial intelligence to reproduce images or voices of candidates to falsely portray them without disclosing that the images or words are fake. do.

The injunction was sought by conservative commentator Christopher Coles, who has created numerous deepfake videos satirizing Democrats, including Vice President Kamala Harris, the party’s presidential candidate. Governor Gavin Newsom said: one of those videos The video, which shows a clip of Harris and a deepfake version of her audio talking about being the “ultimate diversity hire” and professing both ignorance and incompetence, While AB 2839 was signed into law, the measure was actually introduced in February, long before Coles’ Harris video. It became a hot topic in X.

When asked about the sentence in X, Coles said: For now. “

of ruling U.S. District Judge John A. Mendez highlighted the tension between efforts to protect against AI-powered counterfeiting that could sway elections and the Bill of Rights’ strong protections for political speech. exemplifies.

In granting the preliminary injunction, Mendez wrote, “When political speech and electoral politics are at issue, the First Amendment does not require courts to uphold a state’s attempt to suppress speech. “It almost explicitly states that speech should be allowed to flourish…” [M]AB 2839’s Ost functions as a hammer instead of a scalpel, serving as a blunt instrument that prevents expression of humor and unconstitutionally stifles the free and unfettered exchange of ideas so important to American democratic discourse I will. ”

Robert Wiseman, co-chair of Public Citizen, countered that “the First Amendment should not tie our hands to addressing serious, foreseeable, and real threats to our democracy.” .

Robert Wiseman, executive director of Public Citizen, speaks with U.S. companies and businesses to form the Business for Democracy campaign to oppose unlimited corporate spending in public elections, Friday, January 21, 2011, in Washington. He spoke at a press conference that brought together leaders.

(Nick Wass/Associated Press)

Wiseman said 20 states have adopted laws that follow the same basic approach: requiring ads that use AI to manipulate content to be labeled as such. But Wiseman said AB 2839 has unique factors that may have influenced Mendez’s thinking, such as a requirement that disclosures be as large as the largest text that appears in the ad. It is said that there were several.

In his ruling, Mendez noted that the First Amendment also applies to false and misleading speech. Even on a topic as important as protecting elections, he wrote, lawmakers can only regulate speech through the least restrictive means.

AB 2839 requires political videos to continually display necessary disclosures about manipulation, but does not use the most restrictive measures to protect election integrity, Mendez wrote. . A less restrictive approach would be “counter-speech,” he wrote, without explaining what that meant.

Mr. Wiseman responded that “counter-speech is not an adequate remedy.” The problem with deepfakes, he says, is not that they make false claims or innuendos about candidates. “The problem is that candidates are making it look like they’re saying or doing things that they’re not actually doing,” he said. He faces the near-impossible task of explaining what he hasn’t done or said, which he says is much harder than countering false accusations by opponents or accusations by political action committees.

He said requiring disclosure of operations is not a perfect solution to the challenges posed by deepfake ads. However, it is the least restrictive remedy.

lianna keesing No. 1Pro-democracy groups said the creation of deepfakes is not necessarily the problem. “The important thing is that that false and deceptive content is amplified,” said Keesing, the group’s campaign manager.

Alix Fraser, Director of Technology Transformation No. 1said the most important thing lawmakers can do is address how technology platforms are designed. “What are the guardrails around it? There are basically none,” he said, adding, “That’s the central question we think about.”