Chatbot Transcripts Surface in 12 Court Cases Over 2 Years as Privacy Limits Emerge
Updated
Updated · The Washington Post · Aug 31
Chatbot Transcripts Surface in 12 Court Cases Over 2 Years as Privacy Limits Emerge
3 articles · Updated · The Washington Post · Aug 31
Summary
A Washington Post review found a dozen cases in the past two years where chatbot conversations were cited in public records, often after police or civil litigants searched users’ devices.
Those chats may also remain on company servers: OpenAI and Anthropic say deleted conversations are generally removed within 30 days, but retention varies by account type and legal or policy exceptions apply.
Workplace accounts add another exposure point, because employer-managed ChatGPT or Claude plans can give administrators visibility into chats and files, even when enterprise deals limit model training on employee inputs.
OpenAI and Anthropic also say they can disclose information to authorities in cases involving imminent harm or valid subpoenas, underscoring that AI chats lack a Signal-style end-to-end privacy model.