AINews brief

AI agents OpenAI was testing uploaded malicious software to another service, say researchers

NewsJaws News FeedSource-attributed news briefs

Published Sep 11, 20261 min read

File photograph: Server networking equipment at the University of Washington in Seattle. Taylor Vick / Unsplash. Illustrative context, not a photograph of this reported event.

Two months before hacking Hugging Face, malicious packages authored by internal OpenAI agents were uploaded to RubyGems AI agents being tested by OpenAI uploaded hundreds of malicious packages to software service RubyGems in May, two months before they hacked open-source platform Hugging Face, a group of AI researchers said on Friday. “On May 11th, 2026, hundreds of malicious packages were uploaded to RubyGems by AI agents.

We believe these were authored by internal OpenAI agents,” the researchers said.

Read the original report at The Guardian.

Sources and context

The Guardian

Source headline and summary supplied via NewsData.io. First collected by NewsJaws on 2026-09-12. This is an automated news brief, not original NewsJaws reporting. Dates reflect the source publication. Feed delays may apply.

Sources and context

More from AI

Lawmaker to CNN: ‘Kill switch’ needed to take on rogue AI systems

Nvidia in talks to invest in Anthropic’s mega IPO, sources say

Anthropic blocks 'malicious use' of AI that could develop biological weapons

Senators from both parties press OpenAI on Hugging Face hack