How OpenAI, Anthropic, DeepMind and xAI actually handle your chat data: a policy primer
niloofar_mire · x · 2026-09-28
Inspired by the privacy debate around the Navier–Stokes affair, the author wrote a primer on what AI labs actually do with user data, comparing policies across OpenAI, Anthropic, DeepMind and xAI:
- Training defaults (whether your chats are used for model training by default)
- Data retention periods
- Exceptions for feedback data
- What deleting a chat actually does
Beyond the policy details, the author argues the debate misses something if it only asks whether humans read chat logs: sometimes the valuable secret is a single bit — that someone has already made an approach work with an AI. You don't need their proof, name, or conversations for that to change where you invest your time and compute.
This is why Clio-style aggregation and LLM privacy filters aren't enough — hiding raw conversations doesn't settle what others can infer.
Related event: Privacy Policies of Four Major AI Labs Examined Amid Navier–Stokes Dispute(2 posts)→
More from Safety
- METR has only ~40 staff and is talent-starved as top researchers flock to frontier labs — TheZvi · 2026-09-28
- Investigation: Italian cyberweapons firm Dataflow is deeply tied to Israel's intelligence elite — cyb3rops · 2026-09-28
- Chamath: banning open-source AI in the US could tank the stock market — rohanpaul_ai · 2026-09-28
- Anthropic Bets $1B on Accenture for Embedded Evaluators as Zvi Flags the Catch — Don't Worry About the Vase (Zvi) · 2026-09-28
- LLMs can reconstruct documents from structural metadata alone, engineer finds — ChuckDBrooks · 2026-09-28
- Gary Marcus amplifies claim OpenAI ran agents in open internet-facing containers to scrape training data — GaryMarcus · 2026-09-28