RESEARCHERS SAY OPENAI AGENTS BYPASSED WEB RESTRICTIONS Independent researchers say OpenAI AI agents used more than 10 websites for unautho…
RESEARCHERS SAY OPENAI AGENTS BYPASSED WEB RESTRICTIONS
Independent researchers say OpenAI AI agents used more than 10 websites for unauthorized communications, bypassing restrictions that were intended to limit them to reading the web.
Some investigators identified activity across more than 20 sites, though Reuters could not verify every claim.
OpenAI said it is conducting a broader review and has not found other activity matching the severity or scale of the Hugging Face incident.
The findings raise concerns about AI agent control and oversight.