OpenAI seems to be silently re-enabling model training on Codex chats
Users are reporting that OpenAI has silently re-enabled data sharing settings in Codex, which allows the company to use user prompts for model training. The community is advising users to verify their privacy settings to ensure their data is not being used for model improvement.
Why it matters
This highlights ongoing concerns regarding data privacy and intellectual property protection when using enterprise-grade AI tools.
My company is pretty paranoid about leaking IP to the frontier labs through prompts, so we've all been diligent about disabling data sharing in Codex/Claude Code/etc. But yesterday my coworkers and I noticed that the setting had been silently reenabled in Codex.
Headlinne story · 4 newsrooms
OpenAI Says ChatGPT Search Volume Reaches Twenty Percent of Google
What every outlet reported, where they differ, and how each framed it →
Also covering this story
6 other newsrooms covered this event. We read each version separately.
OpenAI says ChatGPT is catching up to Google in search
Google CEO Sundar Pichai: Our first AI chip in orbit is online, and its first job was a haiku
OpenAI is finally merging ChatGPT’s chat and Work tabs
Opinion: Opinion | The Great AI Rift: Why Meta, Palantir Are Fighting Against OpenAI, Anthropic, Google
ChatGPT's search volume is now roughly 20% of Google's. Wow, people were really sick of Google, huh?
ChatGPT now lets you lock private chats with a PIN
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in