Your Conversations with AI Can Be Used Against You in Court

Americans increasingly use artificial intelligence chatbots as advisers, sounding boards, and even confidants. We ask about medical symptoms, relationships, workplace disputes, political ideas, and legal problems. Many of us reveal thoughts we would hesitate to share with another person.

These conversations seem private, but they can become quite public.

The Washington Post found a dozen civil and criminal cases over the past two years in which chatbot conversations surfaced in legal proceedings. As AI systems become more deeply integrated into daily life, such disclosures are likely to become routine.

Civil discovery can allow lawyers to search broadly through an opposing party’s devices and digital records. In one Michigan dispute, a former employee’s questions to ChatGPT about whether Yahoo could recover a deleted email became evidence in litigation brought by his former employer.

In another case, a 15-year-old boy sued several social media companies, alleging that their platforms contributed to years of addiction and mental-health problems. Defense attorneys obtained his ChatGPT conversations, including deeply personal exchanges about his father, and placed them in the public court record.

Mike Morgan, an attorney whose firm represented the teenager, observed that a teen “may type things into a chatbot that he may not say to a therapist, a parent, or a friend.” Yet that teenager may have no idea his words could later appear in an expert report or public court filing.

Relevant evidence is not immune from discovery merely because it was shared with a chatbot. Courts may permit the use of such conversations as evidence of fraud, the destruction of evidence, and damaging admissions.

Those who don’t understand this risk having their deepest secrets exposed in court.

Chatbot records may contain a uniquely revealing mixture of private facts, half-formed ideas, emotional venting, political and religious questions, and hypothetical scenarios. Michael Price, litigation director of the National Association of Criminal Defense Lawyers’ Fourth Amendment Center, aptly calls these conversations “a window into the soul to reveal the privacies of life.”

Nevertheless, chatbot conversations, like diaries or internet searches, have no immunity from legal proceedings, whatever the implications for privacy and First Amendment expression. Asking a chatbot a legal, psychological, or theological question does not create an attorney-client, therapist-patient, or clergy privilege.

Anything you tell an AI chatbot might someday be demanded by an employer, business adversary, divorce lawyer, or ideological opponent.

As more people understand this, they might censor their inquiries. Their freedom to investigate and privately test ideas might be chilled. At the very least, we hope courts will require opposing counsel’s requests for chatbot histories to be specific, relevant, and proportional to the needs of a case. Judges should seal intensely personal information – especially records involving minors.
​
But as long as a court relies on due process – and in criminal cases, a warrant – your chatbot inquiries can be made public in court.

DONATE & HELP US DEFEND YOUR FIRST AMENDMENT RIGHTS