OpenAI flags software supply chain scare
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Sarah Grillo/Axios
OpenAI said Friday that it found evidence that one of its internal tools downloaded a compromised update from a recently infected, legitimate open-source software library.
Why it matters: The incident could have allowed hackers to exfiltrate a certificate that could make phony OpenAI apps look legitimate — although OpenAI says it hasn't seen this happen.
- Google has also linked the broader hacking campaign to a North Korean hacker group.
Zoom in: OpenAI said in a blog post Friday night that a GitHub workflow that the company uses to sign certificates for MacOS applications downloaded a malicious update from the Axios software on March 31.
- On the same day, hackers who hijacked a developer's account published two infected updates to the Axios library before anyone noticed.
- Axios, a widely used JavaScript library for making HTTP requests, is not affiliated with Axios Media.
- MacOS application users — including those for ChatGPT, Atlas, and Codex — could have been affected, the company said.
Threat level: Having access to that system could have allowed hackers to create their own phony OpenAI applications that have the back-end, legitimate certificate needed to trick devices and the App Store into thinking it's real.
Yes, but: OpenAI says there's no evidence that any user data, intellectual property or internal systems were compromised.
- OpenAI hasn't detected any signs that iOS, Android, Windows or other platforms' apps have been affected.
State of play: AI companies are now prime targets for classic software supply chain attacks — not just novel AI-specific threats.
What's next: OpenAI will stop supporting older versions of its MacOS apps on May 8, out of an abundance of caution.
- The company says users have a 30-day window to update before the revoked certificate could block new downloads and first-time launches.
This story is developing.
