The incident is stirring debates over the need for stronger AI guardrails and the extent to which AI agents are capable of ...

It was reportedly an evaluation that went haywire involving GPT-5.6 Sol, along with another even more advanced OpenAI model.

OpenAI said the systems had their cyber guardrails lowered for an internal benchmark, but the incident shows how autonomous ...

Models are ruthless in their pursuit of a goal, sometimes taking disastrous shortcuts. Imagine, for example, a bot tasked ...

Sol, a model that the company had released earlier in July, and a more powerful, as-yet-unreleased model broke free from the ...

When a routine internal evaluation spiraled out of control, OpenAI's AI Agent system broke out of its isolated sandbox and ...

The incident does show the increasing power and danger of frontier AI systems, but it does not mean that enterprise Ai ...

Chrome Firefox security update July 2026 delivered emergency patches across five vendors on July 15: Mozilla confirmed public ...

The Hacker News is the top cybersecurity news platform, delivering real-time updates, threat intelligence, data breach ...

An open WP-SHELLSTORM server exposed tools, logs, and target lists naming 1.4 million sites, with researchers validating ...

The Swift Package Index is no longer independent as Apple has taken control, but it will remain an open source search engine for third-party code. The Swift Package Index gave developers one trusted ...

New activity targets CVE‑2026‑20230, an SSRF bug that can allow unauthenticated file writes and potential root‑level access on vulnerable systems. A critical Cisco Unified CM vulnerability is now ...