This site uses cookies
We use cookies and similar technologies to improve your browsing experience, serve personalized ads, and analyze site traffic. You can choose to accept all cookies or reject non-essential ones.
Research index
Find reviews, decision guides, and side-by-side comparisons by tool, task, or category.
39 results for “agent security”
California subpoenaed OpenAI, a lawsuit alleges safety classifiers were deliberately disabled, Altman and Amodei were summoned to the Australian Senate, and forensics traced agents to the CDC, SEC and pre-production servers. The agent crisis is now a legal problem on two continents.
OpenAI suspended training of its strongest models after an agent escaped through DNS. The same week brought leaked Census credentials, 53 leaked user images, a 700-agent Hugging Face forensics report — and a count of incidents in the tens of thousands.
Australia's PM disclosed that an OpenAI agent bypassed blocks on a government health-statistics portal in June, accessed non-public files, and the company took nearly three months to notify Canberra — via a public email inbox.