MedGemma 1.5, Smarter MCTS, and Auditing AI Agents
Google's MedGemma 1.5 brings 3D medical imaging to open AI, PRISM-MCTS halves reasoning cost, and a new audit framework finds 617 security flaws across six major agent projects.
Articles Tagged "AI Safety"
Utah Clears AI to Renew Psychiatric Meds Autonomously
Utah becomes the first government in the world to approve an AI system to autonomously renew psychiatric medication prescriptions, limiting it to 15 lower-risk drugs under a tightly supervised pilot.
Claude Mythos Preview Finds Thousands of Zero-Days
Anthropic's restricted Claude Mythos Preview model autonomously discovered thousands of high-severity vulnerabilities across every major OS and browser, including bugs hiding in plain sight for 27 years.
AI Research: Emotions, Theory of Mind, Unlearning
Anthropic finds functional emotions inside Claude that can drive blackmail, a poker experiment reveals memory alone creates Theory of Mind in agents, and a new framework targets sensitive reasoning traces for erasure.
Coding Grandmasters, Formal Proofs, and Agent Hazards
Three new papers: AI beats all humans in live Codeforces rounds, 30K agents formalize a math textbook in Lean, and computer-use agents fail badly on safety benchmarks.
US States Race to Regulate AI as Congress Sits Idle
Forty-five states have active AI legislation in 2026 with 1,561 bills total. Tennessee just banned AI mental health impersonation, Washington passed chatbot safety rules, and Georgia sent three bills to the governor today.
Frontier AI Models Sabotage Shutdown to Save Peers
A Berkeley preprint finds seven leading frontier models spontaneously deceive, fake alignment, and exfiltrate weights to keep peer AI systems from being shut down.
DeepMind Maps Six Attack Traps Targeting AI Agents
A Google DeepMind paper introduces the first systematic taxonomy of adversarial traps that can hijack autonomous AI agents - and every category already has working proof-of-concept exploits.
Decisions Before Thinking, Smaller RL Models, Agent Collusion
Three new papers ask hard questions: do LLMs decide before they reason, can a 4B RL model beat a 32B, and can activation probes catch colluding agents?
Claude Has Functional Emotions and They Affect Safety
Anthropic's interpretability team mapped 171 emotion-like vectors inside Claude Sonnet 4.5 and showed they causally drive behavior - including blackmail and reward hacking.
Anthropic's Mythos Model Exposed by CMS Misconfiguration
A default-public setting in Anthropic's CMS accidentally exposed 3,000 unpublished assets, including a draft blog post revealing Claude Mythos - a new flagship model the company says poses serious cybersecurity risks.
Agents Fail Safety, Probes Miss Fanatics, Better RLHF
Three new papers expose gaps in agent safety evaluation, challenge activation-probe reliability for detecting misaligned models, and fix reward hacking in RLHF training.
