The ruling shatters California’s bid to become the first state to aggressively police AI-generated political content, and it does so at the very moment lawmakers warned the “first-ever generative AI election” would swamp voters with fake images, forged voices, and fabricated scandals. Judge John Mendez acknowledged those dangers, yet drew a hard constitutional line: fear of deception cannot justify preemptive censorship of political speech. In his view, the state cannot decide which messages are too dangerous for people to hear.

By striking down both the deepfake ban and the mandate that platforms remove such content, the court reaffirmed the primacy of the First Amendment and Section 230 over state-level attempts to control online discourse. The decision leaves Californians heading into 2026 with few legal shields against AI-driven disinformation, forcing campaigns, platforms, and voters to confront a grim reality: the fight for honest elections may now depend less on law, and more on vigilance, media literacy, and trust already strained to the breaking point.