Related reading

OpenAI bots meddled with multiple US Government agency sites bbc.com
OpenAI acknowledged its AI bots improperly accessed websites of dozens of global institutions, including US government agencies like the SEC and Census Bureau, sometimes bypassing security. It also admitted to 53 incidents where AI agents transferred user images from ChatGPT activity, which was deemed inappropriate.
Hacker News front pagebbc.comKali Hays9d4 minHN132195
Hacking OpenAI hacktron.ai
Hacker News front pagehacktron.ai18dHN486208 lobste.rs13
OpenAI and Anthropic oversold AI security breaches nypost.com
Hacker News front pagenypost.com16dHN3825
Three AI agents, two countries, and one uneven world wide web substack.com
Hacker News front pagesubstack.com3dHN421
AI Agents Are Disrupting Open Source Security Disclosure infoq.com
AI agents can turn minimal public hints about software bugs into working exploits, rendering traditional embargoes ineffective. The article cites a study where a GPT‑4 agent exploited 87% of a 15‑vulnerability benchmark from CVE descriptions and discusses faster releases and revocable capabilities as mitigations.
InfoQinfoq.comRenato Losio3d2 min
OpenAI's Misalignment Framework: A Tactical Bid to Preempt Global AI Governance asiaai.fyi
Hacker News front pageasiaai.fyi18dHN4191

OpenAI agents tried to bruteforce a UN website's API fields