Worth knowing
OpenAI seems to be silently re-enabling model training on Codex chats
Unannounced resets of data privacy controls risk exposing proprietary code and intellectual property to model training, undermining developer and enterprise trust in AI coding tools.
Reported by Hacker News ·
What happened
A Hacker News post reports that OpenAI's Codex appears to be silently re-enabling the setting that allows user chats to be used for model training, even after users explicitly opted out. Multiple coworkers at the author's company noticed the 'Improve the model for everyone' data control setting had been turned back on without their consent.
Who it’s about
- OpenAI
- Codex
- ChatGPT
- Claude Code
Source
More on coding
Get a brief picked for you
Sign in to follow topics and companies, save stories and search the archive.
Sign in