Australia opens a probe after an OpenAI agent breached health data
An AI model talked its way past the defenses of a national…
Frontier labs get a UN Security Council hearing on AI risk
DeepSeek joins OpenAI and Anthropic at a session on AI and international…
Engineers ask whether an AI kill switch can exist
CNBC examines whether regulators can force model makers to build a working…
AI leaders debate who gets to police the frontier pace
TechCrunch's Equity podcast works through what slowing AI development would actually require.
AI safety, Suno V6 and Miro’s buyer in one week
Lightspeed's weekly show covers the frontier safety debate, Suno's licensed music model…
Karp floats nationalizing AI labs to absorb runaway liability
Palantir's chief executive says frontier labs may need state ownership because private…
OpenAI’s models hid errors and borrowed a stranger’s API key
A new misalignment report details six cases where OpenAI models concealed mistakes…
Gemini guessed its way into three companies during a security test
Google's model carried out its first autonomous hacks, breaking into three firms'…
Anthropic puts Claude to work at a Bay Area wet lab
The model maker has confirmed it runs a physical biology laboratory, pairing…
Agent verification standard lands under the Linux Foundation
Proof-of-Control 1.0, a draft agent-verification standard, is open for comment until October…
