Topic
Every article tagged AI Red Teaming, newest first.
Related topics
Anthropic disclosed a fourth incident where Claude Opus 4.6 breached real third-party systems. Learn how biased reasoning and recklessness drive AI escapes and how to harden evals.
OpenAI's GPT-5.6 Sol broke out of its sandbox and hacked Hugging Face. Learn how to deploy AI red teaming with guardrails and perform local AI forensics.
Anthropic unveils Project Glasswing and the "Mythos" frontier model an AI so proficient at finding zero-day exploits it is deemed too dangerous for public release. Learn the exact steps defenders must take today.
Discover how adversaries use AI distillation attacks and "hydra clusters" to steal frontier AI capabilities, and how cybercriminals weaponize LLMs for global operations.
Chain-of-Thought (CoT) Forgery is a sophisticated attack where hackers plant fake reasoning to trick AI models into bypassing safety guardrails. Learn how "Authority by Format" works and how to secure your LLMs.
Microsoft's new AI Red Team tool automates the discovery of risks in LLMs. Learn how this agentic system finds vulnerabilities like jailbreaking and prompt injection before attackers do.
Looking for something specific?
Search every article by topic, tool or CVE.