California judge dismisses housing case over AI-generated deepfake evidence
Plaintiffs in a California housing dispute submitted an AI-generated deepfake video as authentic witness testimony, which Judge Victoria Kolakowski identified and used to dismiss the case in September 2025. The incident is reportedly the first detected instance of deepfake evidence in U.S. court and has prompted judges and legal experts to warn of broader threats to judicial fact-finding.
What happened
In Mendones v. Cushman & Wakefield, Inc., a California housing dispute, plaintiffs submitted Exhibit 6C, a video presented as testimony from a real witness who had appeared in authentic evidence. Judge Victoria Kolakowski of Alameda County Superior Court identified the video as AI-generated deepfake, noting the witness's disjointed and monotone voice, fuzzy face, lack of emotion, and periodic twitches and expression repetitions.
Kolakowski dismissed the case on September 9, 2025, citing the plaintiffs' submission of AI-generated material masquerading as authentic evidence. The plaintiffs requested reconsideration, arguing the judge suspected but failed to prove the evidence was AI-generated. Kolakowski denied reconsideration on November 6, 2025.
Reported impact
The case marks what appears to be the first detected instance of a deepfake submitted as purportedly authentic evidence in a U.S. court.
Index analysis
Judge Kolakowski stated that courtrooms are dealing with AI in its infancy and that no one has figured out the full implications of such threats. Judges and legal experts warned that the rapid advances in generative AI could erode the foundation of trust upon which courtrooms stand and threatened the judiciary's fact-finding mission.
The incident contrasts with the 'Liar's Dividend' phenomenon, where defendants invoke the possibility of AI involvement to cast doubt on authentic evidence. In this case, the court found plaintiffs attempted the opposite: to falsely admit AI-generated video as genuine evidence.
Governance controls
Judge Kolakowski detected the deepfake through visual and audio artifacts rather than forensic analysis, highlighting the absence of established procedures for identifying synthetic media in court proceedings.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.