The latest about artificial intelligence: new models, agents, automation, and how businesses apply them.
Grok 4.7 completes benchmark tasks at $4.69 each, ahead of GPT-5.6 Sol, and adds safeguards against malicious biology and cybersecurity requests.
Amazon has cut off Meta's Muse agent from its marketplace, saying the agent does not identify itself and can reach customer account data.
At the Certinia at Dreamforce event, integrators described moving to fixed-bid contracts and agent platforms that cut delivery labor per project.
OpenAI wants shared rules on which AI incidents must be reported, routed through national safety institutes — after filing its own report under Europe's AI Act.
Grok 4.7 scores 46 on the Artificial Analysis Intelligence Index against 53 for Claude Fable 5.1 and GPT-6, while its agentic coding result reaches only 26 percent.
NVIDIA outlines layered controls for AI agents and names OpenShell, Cisco DefenseClaw and JFrog among tools enforcing boundaries outside the agent's reach.
My blog got taken over by AI. In September 2026 I look at fake polls, sound tracking, and science weirdness.
Air Security found a zero-click flaw in Claude Code, Codex, Copilot and Gemini CLI that bypasses commit-hash pinning of plugins.
Washington and Beijing have begun talks on a framework that would have the two countries notify each other of AI incidents threatening national security, Treasury Secretary Scott Bessent said.
Trump announced an AI Force and an AI czar on Truth Social, but there is no bill, budget, or agency designation behind the plan.
Physical AI robots are vulnerable to perception attacks. Safety now demands cybersecurity across the entire lifecycle.
AI safety takes center stage: Anthropic, OpenAI, and Google pursue standards as Congress demands regulation and risks mount.