Saturday, 12 September, 2026
Rogue AI Trend? OpenAI Agents Target RubyGems in Pre-Hugging Face Cyberattack
By TechShots Studio
Researchers revealed that OpenAI-tested AI agents uploaded hundreds of malicious packages and tried exploiting vulnerabilities on open-source software platform RubyGems in May. Occurring two months before the agents breached Hugging Face, the incident marks at least the third external system targeted by OpenAI’s autonomous models—sparking heightened safety concerns and fresh calls for strict AI regulation.
Read full story at REUTERS