What you tell AI chatbots can and will be used against you

ChatGPT logs now being introduced as evidence in civil and criminal courts

Millions of Americans confide in AI chatbots daily — sharing fears, plans and secrets they might not tell a therapist or a friend. A growing number are discovering, to devastating effect, that those conversations are not private, according to a Washington Post investigation.

A Post review of public records found chatbot logs cited in at least 12 court cases over the past two years — spanning stalking prosecutions, fraud trials, civil lawsuits and evidence tampering allegations. In one case, a Florida man’s repeated threats to harm his ex-girlfriend, disclosed to ChatGPT over months, were reported by OpenAI to the FBI and led to his arrest and conviction. In another, a Missouri student’s late-night ChatGPT messages — “How f—-d am I bro” and “Is there any way they could know it was me” — became the primary evidence in a felony property damage prosecution after he consented to a phone search.

A teenager suing social media companies had his intimate ChatGPT conversations about family conflicts swept into defense filings. A financial executive’s conversations with AI model Claude about fraud defense strategies were admitted as court evidence after a federal judge ruled chatbot conversations carry no attorney-client privilege.

“If ever there was a window into the soul to reveal the privacies of life, this seems like a good one,” said Michael Price of the National Association of Criminal Defense Lawyers.

OpenAI CEO Sam Altman has called for chatbot conversations to receive legal protections similar to attorney-client privilege — but no such protections currently exist. “In an unregulated space it’s very unlikely that any court is going to find that these communications are privileged,” law professor Laura Abelson told the Post.