OpenAI's ChatGPT share feature with a 'Make this chat discoverable' option led to over 100,000 private conversations being indexed by Google and permanently archived, exposing sensitive personal information, business plans, API keys, and confessions.
OpenAI introduced a share feature for ChatGPT conversations that included a checkbox labeled 'Make this chat discoverable' which allowed search engines to index the shared conversations. Users who clicked this option unknowingly made their private ChatGPT conversations publicly searchable on Google. The incident exposed over 100,000 conversations containing sensitive information including resumes, job applications, internal business plans, API keys, developer prompts, personal confessions, and academic fraud documentation. Researchers found conversations revealing corporate strategies to displace indigenous communities, political dissent in authoritarian regimes, and systematic academic misconduct. While OpenAI quickly responded by adding noindex and nofollow tags and requesting Google remove the indexed pages, over 100,000 conversations had already been permanently archived by Archive.org's Wayback Machine. The incident affected multiple AI platforms beyond ChatGPT, including Claude, Copilot, Grok, Mistral, and Qwen, with researchers discovering 143,000 total shared conversations across all platforms. OpenAI has not requested removal of the archived conversations from Archive.org, leaving them permanently accessible.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that memorize and leak sensitive personal data or infer private information about individuals without their consent. Unexpected or unauthorized sharing of data and information can compromise user expectation of privacy, assist identity theft, or cause loss of confidential intellectual property.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed