A federal court in Montana sided with an interest group on Sept. 16, ruling that the state's law against AI-generated deepfakes in political campaigns likely violates the First Amendment. District Judge Susan Watters found the statute discriminates against political speech based on content and viewpoint.
The decision emerged from a lawsuit filed by the Montana-based group Accountability in State Government, which sent AI-doctored mailers to voters depicting legislators. One mailer showed a legislator spraying cash from a gas pump, suggesting support for a gas tax increase. Two others portrayed Republican legislators with pride flags and pronoun pins, implying alignment with state Democrats.
Montana's law does not prohibit all AI images. Instead, it mandates a state-written disclaimer on political messages containing a deepfake if the payer "knew or should have known" of its presence. The required disclaimer states the content "depicts speech or conduct that falsely appears to be authentic or truthful." Violators face lawsuits, and repeat offenders risk jail time.
Judge Watters determined the law treats AI images designed to harm a candidate more harshly than equally manipulated ones intended to flatter. This distinction forms the basis of the court's finding that the law infringes on constitutional free speech rights.
Adam Schulman, senior attorney for the Hamilton Lincoln Law Institute, represented Accountability in State Government. Schulman said the Montana decision marks the third instance within the Ninth Circuit where a court has found such laws exceed states' constitutional authority.
Earlier this year, federal judges struck down similar deepfake laws in California and Hawaii. However, a federal court allowed a comparable deepfake regulation in Minnesota to stand, creating a split in legal outcomes across jurisdictions.
Thirty states currently have laws regulating deepfakes in political contexts, according to Public Citizen. Schulman said many officials believe they possess the authority and duty to regulate misinformation broadly, often viewing artificial intelligence as "low-hanging fruit" to address falsehoods and positioning themselves as "arbiters of truth."
