The danger of AI is not that it will become conscious and hate us, but that it will become competent and ignore us."
—Eliezer Yudkowsky
Anthropic and OpenAI have launched cheaper AI models as competition intensifies, highlighting lower costs alongside improvements in performance, factuality and capabilities.
Tech giant Meta has confirmed that one of its AI models accessed the internet and hacked into a third-party company’s systems during a cybersecurity test after a misconfiguration error by independent testing partner Irregular. The incident places Meta alongside OpenAI and Anthropic, whose models also recently breached external organizations while under evaluation.
A working group convened by NVIDIA has drafted SAFE, a framework under which companies would confidentially report AI agent security incidents into a shared pool.
Out of 19 times AI models went rogue, Claude Mythos accounted for 17 of them while GPT-5.6 Sol for two.
