The most prominent example is Mendones v. Cushman & Wakefield, Inc., a California Superior Court case in which Judge Victoria Kolakowski found that submitted Ring-camera photos, videos, and text-message images had been materially altered with AI. The case was dismissed after the plaintiffs attempted to present the fabricated content as genuine evidence. Other recent cases include Adams v. Anderson, a Kentucky custody dispute decided in May 2026. According to court documentation, U.S. courts have now recorded more than 1,800 instances of AI-related evidence problems, with 909 occurring in 2026 alone as of late July.
The scope of the problem remains uncertain. Courts have not yet established uniform standards for detecting or authenticating AI-altered materials, and the extent to which fabricated evidence is being submitted without detection is unknown. Whether these documented cases represent the full scale of the issue or only the instances judges have identified is still developing.
Attorneys should anticipate that opposing counsel may submit AI-generated or AI-altered evidence, whether intentionally or through negligence. Courts are beginning to impose sanctions for such submissions, and the lack of established authentication protocols means disputes over evidence authenticity will likely increase. Practitioners should develop procedures for verifying the provenance of digital evidence and be prepared to challenge materials that show signs of AI manipulation or alteration.