Woman Jailed After Court Accepted AI-Generated Text as Evidence

A woman was convicted and jailed based on text messages that a court accepted as evidence without verification. The messages were generated by an AI system. The system had a documented process with multiple steps. None of those steps included verification of accuracy or authenticity. The checklist was completed. The checkbox for "is this real" does not exist on the form.

Courts have always relied on evidentiary procedures. Those procedures assume human testimony or documentary proof. Both can be cross-examined or authenticated. An AI-generated text has no author in the traditional sense. The process that produced it is deterministic but opaque. Many jurisdictions have not updated their rules of evidence. The rules of evidence have not been updated.

This will happen again. Defense attorneys will begin requesting AI verification as a matter of course. Some will succeed. Some will not. The woman remains in custody. The process that convicted her remains unchanged.