OpenAI has filed a legal objection to a New York Times subpoena seeking access to 20 million private ChatGPT conversations. The media outlet’s demand, filed in a New York federal court, would give it a snapshot of user interactions spanning several months. OpenAI says the request exceeds any legitimate journalistic need and threatens user confidentiality. In parallel, the company is accelerating rollout of new security features, including stricter data encryption at rest and tighter access controls for internal reviewers.
The move matters because it tests the limits of journalistic subpoenas in the AI age. If courts allow the NYT to obtain raw chat data, developers could face new pressures to hand over user‑generated content, potentially chilling adoption. OpenAI’s pre‑emptive technical hardening also signals that firms are preparing for a broader wave of privacy litigation, something rivals like Anthropic and Google have already hinted at in their own policy updates.
For now, the dispute adds another chapter to the growing tug‑of‑war between data‑driven AI services and traditional media’s investigative tools, reminding regulators that user privacy is becoming a battlefield rather than a footnote.