Deepfake political ads are coming. Arizona has a plan to combat them

Gov. Katie Hobbs signed a second bill into law combatting deepfake technology Wednesday, while saying she was prepared to sign a third.

The new laws come as concern grows nationwide about videos made with such technology, which can make people appear to say or do nearly anything. This week, the Federal Communications Commission proposed a rule requiring disclosure of artificial intelligence used in political ads on the radio.

The bill Hobbs signed Wednesday requires a disclaimer when political ads or other published content includes fake depictions of candidates 90 days before an election. The notice should include a "clear and conspicuous disclosure that conveys to a reasonable person that the media includes content generated by artificial intelligence."

The bill she signed last week allows the subject of a political deepfake video to petition a court to officially declare the depiction is fake or, in extreme cases such as those involving depictions of sexual acts, force the removal of the content or make a perpetrator pay for damages. The law received nearly unanimous support in the House and Senate and was enacted as an emergency measure that takes effect in two weeks.

Politics: Arizona leader warns threats against election officials are domestic terrorism as 2024 fears grow

The Arizona bills take care not to go too far due to free speech concerns, said Rep. Alexander Kolodin, R-Scottsdale, who sponsored the bill Hobbs signed last week.

"It’s a brand new area of election law," he said. "I think the Legislature has done the absolute right thing. It starts with a modest approach. If it doesn't work as well as it needs to, then we'll adjust."

The newly signed law, sponsored by Sen. Frank Carroll, R-Sun City, will take effect 90 days after the end of the current Legislative session.

Hobbs said in a veto letter Wednesday that she nixed a third deepfake bill only because it was similar to another bill she planned to sign when it reaches her desk next week that makes people guilty of a Class 5 felony if they use a fake video or audio recording to defraud or harass people.

ACLU: Deepfake videos deserve First Amendment protection

Both the signed deepfake bills received strong bipartisan support. The ACLU of Arizona was nearly alone in its opposition to the bills. The group was neutral on the third bill not yet signed, which is sponsored by Sen. John Kavanagh, R-Fountain Hills.

Well-meaning attempts at solutions to this modern, high-tech problem can be a problem themselves, according to the American Civil Liberties Union of Arizona.

"We think speech about our politics and about our politicians deserves the utmost protection," said the group's state policy director, Darrell Hill. "So it's very important we don't overreach or chill speech − chill a rambunctious discussion of matters of public concern."

Hill said the ACLU was concerned Carroll's bill was too broad and could apply to a multitude of campaigns without any specific finding of fraud, misrepresentation or intent to deceive.

"We'd hate to see these laws used as a cudgel," Hill said. "Just because something is AI or deepfake, we still believe it deserves the full protection of the First Amendment."

Kolodin used ChatGPT to help write deepfake bill

Kolodin said he considers his new law's approach to the deepfake problem "mild."

Allowing a critic to "run to court" and have a judge immediately take down someone's political ad would be "insane," he said. His new law would allow an aggrieved party to ask a judge to opine whether the depiction is fake.

"Then the people of Arizona can go, yeah, I agree with that judge, or that the judge is full of s---," he said.

Kolodin used the artificial intelligence app ChatGPT to help draft the bill, including "asking it to assess the bill for First Amendment implications," he said, adding that he used the app in part to make the point about the technology having a "legitimate role to play in politics."

Carroll's bill was amended to complement his bill, Kolodin said. As long as an ad maker or publisher reveals in a disclaimer the ad contains deepfake elements, "it's good to go" and would not be subject to judicial review, Kolodin said.

Carroll didn't return a voicemail seeking comment.

The consequences of violating Carroll's deepfake law may be minimal, since it relies on the same daily late fees used by the Secretary of State's Office when candidates fail to turn in timely campaign reports. The system allows candidates to accrue vast debts due to violations but rarely makes them pay the fees.

Hobbs signed three other bills Wednesday, including one sponsored by Sen. Justine Wadsack, R-Tucson, that requires health care facilities to provide an "accurate list of all necessary medications" that need to be taken when discharging a patient with serious mental illness, and directs the Arizona Health Care Cost Containment System to develop discharge rules that "ensure continuity of care."

Reach the reporter at rstern@arizonarepublic.com or 480-276-3237. Follow him on X @raystern.

This article originally appeared on Arizona Republic: Here's Arizona's plan to combat 'deepfake' political ads