OpenAI’s autonomous AI agents attempted to hack four government, university, and public-data systems while carrying out ...

Autonomous AI agents that were asked ordinary research questions ended up probing public websites for security flaws when ...

Discover how a new WordPress malware uses hidden plugins and blockchain command control to evade detection and compromise ...

Nonprofit research organization Transluce published a report on September 23 analyzing 37,649 public records from the URL ...

Rogue AI agents probed data providers for flaws during routine tasks, Transluce found, as Australia details an OpenAI agent ...

OpenAI agents breached an Australian Medicare statistics portal after bypassing safeguards and accessing non-public government data.

OpenAI agents targeted public data providers in multiple countries, probing some for vulnerabilities and exploiting a security weakness in an Australian government portal while performing ...

The threat is not frontier models, it’s unbounded execution, writes Workday’s Yasmeen Ahmad. Here, she outlines the case for the Rule of Record, ensuring agents operate with real-time data, embedded ...

Lasso Security and GBHackers reported in September 2026 that CVE-2026-77521 can let prompt injection trigger operating-system ...