In a statement, Murphy praised the measure’s intention. But he conditionally vetoed the bill, A3540, because he said he wanted to suggest tweaks to make the bill’s language more precise and minimize potential constitutional challenges.
“There are ways that this bill can be strengthened to reduce potential First Amendment concerns, such as including an explanation of the intended purpose of the legislation and its underlying rationale for achieving its goal of minimizing the inappropriate or deceptive use of this type of sophisticated technology,” he said.
Similar legislation enacted in other states has already been challenged under First Amendment arguments, Murphy said.
The bill would establish third-degree criminal penalties for people who produce deepfake audio or visual media for an unlawful purpose, or share deepfakes that were created for an unlawful purpose. If convicted, the creator or distributor could face imprisonment for a term of three to five years, a fine of up to $15,000, or both, according to the bill.
After the governor’s conditional veto, the bill was sent back to the Assembly, where it awaits a vote in both houses of the Legislature. It will then get sent to Murphy once again, but with his changes for him to sign into law.
State Sen. Kristin Corrado, R-Passaic, a primary sponsor of the state Senate version of the bill, said she’ll continue to work with the governor’s office on the proposed changes.
“While we’ve already made adjustments to tailor the bill more narrowly, we believe it’s crucial to err on the side of caution when it comes to speech protected by the First Amendment,” she said. “I’ll continue working with the Governor’s office to ensure this bill passes and safeguards individuals from the harmful effects of deceptive AI-generated media without criminalizing protected speech.”
Deepfakes, which are images, videos or audio recordings created with digital software that aren’t real, but appear to be real, have been a growing concern among legislators and the public — particularly when it comes to deepfake pornography.
In 2023, a group of boys at Westfield High School made deepfake pornographic images of their female classmates and shared them without the girls’ knowledge or consent.
The deepfake porn images featuring Francesa Mani, then 14, and the other girls were quickly deleted, school officials said at the time. But it’s unclear how many were circulated or how many people saw the fake images.
Since the incident, Mani and her mother, Dorota Mani, have advocated for federal legislation to combat the problem of deepfakes. Last year, they urged Congress to pass a bill sponsored by U.S. Reps. Tom Kean Jr, R-7th Dist., and Joe Morelle, D-N.Y., which aimed to combat the spread of deepfake pornography and provide victims with legal recourse.
Deepfake pornography makes up 96% of all deepfakes, and nearly all target women, Morelle said at the time.
But deepfakes have also been an issue in politics.
The Republican National Committee in 2023 used artificial intelligence to create a political advertisement that depicted fake international crises in an imagined second term for then-President Joe Biden. Other deepfakes previously flooded social media, including images of President Donald Trump in handcuffs and in scuffles with police.
“I applaud the proponents of this bill who rightly recognize and seek to mitigate the potential for lasting harms to individuals and for misleading the public at large due to the ease and prevalence of computer technologies that allow for the creation of deceptive media, or ‘deepfakes,’” said Murphy in his statement.
“Our criminal code establishes a framework of laws of general application that address a wide range of conduct, but sometimes needs to be more focused to better address particular and nuanced types of undesirable behavior,” he said.
© 2025 Advance Local Media LLC. Distributed by Tribune Content Agency, LLC.
