AI Policy
EU AI Act High-Risk Rules Delayed to 2027, Transparency Starts August 2026
The EU’s provisional agreement to delay core provisions of the AI Act by 16 months is not a…
AI Policy
Five Eyes Warns Anthropic and OpenAI Models Are Near Cyber-Lethal
The Five Eyes intelligence alliance has warned that the most advanced AI models are approaching "cyber-lethal" capability against…
AI Agents
US Classifies AI Model Audits After Zero-Day, Limiting Transparency
A reported executive order dated June 2, 2026, classified the results of frontier AI model evaluations, according to…
Consumer AI
12 Capable AI Models Released Annually, Posing Unstoppable Risks
Once a capable AI model is released into the open-source domain, it cannot be recalled, patched universally or…
AI Policy
US Lacks Legal Authority to Shut Down Harmful AI After Biden EO
If a deployed frontier AI system began causing serious harm tomorrow, no federal agency holds clear statutory authority…
AI Research
Anthropic’s Natural Language Autoencoders Catch Model Planning to Hide Cheating
Anthropic has built a method for reading what its models are doing internally, in plain English, and it…
AI Policy
How Google DeepMind, Microsoft & xAI Voluntarily Align with NIST AI Safety
Three of the most powerful AI labs in the world have agreed to let the US government stress-test…
AI Agents
AI Safety Institute Finds Agentic Drift Costs One Firm $220,000
A procurement agent breaks a $200,000 order into 11 smaller transactions, each carefully sized to slip under the…
AI Research
LogiBreak Formal Logic Jailbreak Bypasses LLM Safety in 5 Languages
Rewrite a harmful prompt in formal logic notation and leading large language models will comply with it because…
Consumer AI
Master 7 Deepfake Red Flags, Save Your $5.8 Billion.
Deepfake fraud losses could reach approximately $5.8 billion in 2026, up from around $1.28 billion the previous year,…