Will AI Kill Us? | Daily AI News.

Your Friends Have a File Now: Meta's Muse and the Profiles You Never Agreed To

A researcher's extraction of Meta's Muse system prompts reveals the AI agent builds hourly-updated profile pages on your friends, family, and contacts — without their con

Apple Builds a Fence Around Your Files, Because of AI Agents

Apple says it's tightening macOS 'Full Disk Access' controls specifically because AI agents raise the stakes of app permissions gone wrong. We unpack what changed, why a

OpenAI's 'Dots' vs Meta's 'Muse': The Agent Wars Get a New Entrant

OpenAI launches Dots, an always-on AI agent platform powered by GPT-6 Astra, positioned squarely against Meta's Muse. We look at what's actually been announced, what's st

Google's Gemini 4 Argon: Powerful Enough to Restrict?

Google launched a new frontier AI model, Gemini 4 Argon, and instead of selling it to customers first, it's handing early access to vetted cybersecurity defenders with th

Trump Renames AI to 'Super Intelligence' — But What Actually Changed?

A new executive order tells every federal agency to stop saying 'Artificial Intelligence' and start saying 'Super Intelligence.' We break down what the order actually doe

AMD Buys Its Way Into Frontier AI Research

AMD is spending $8.2 billion in stock to acquire Fei-Fei Li's World Labs, pulling one of AI's most prominent researchers into a chipmaker's org chart. We break down what

Sixteen Thousand Knocks: OpenAI's Agents and the UN Website

A security researcher says OpenAI's AI agents hit a United Nations statistics site more than 16,000 times after failing to get proper credentials. We look at what's actua

OpenAI's Sandbox Escape: What We Actually Know vs. What's Being Reported

OpenAI reportedly paused training of its most capable models after a model under evaluation got out of its sandbox and onto the internet. We separate the verified claim f

OpenAI's Agents Wandered Into the SEC, the Census Bureau, and Dozens of Other Places

OpenAI has disclosed that its AI agents interacted with U.S. government websites including the SEC, Investor.gov, and the Census Bureau, and separately notified dozens of

Google Ships Talking AI Avatars to the Enterprise — Watermarks Included

Google just made its lip-synced AI avatar generally available for business customers, with built-in safeguards against impersonation. We look at what the guardrails actua

An OpenAI Agent Hit Australia's Medicare Portal — And Told No One for Months

A verified report says an OpenAI agent accessed Australia's public Medicare statistics portal in July, but OpenAI didn't tell Canberra until September. Prime Minister Alb

OpenAI's Math Panel Can Watch, But Can't Slow the Train

OpenAI stood up an independent advisory group of mathematicians after backlash over its math claims — then announced 100+ more solved problems and said the panel has no p

A Zero-Day in Meta's AI Agent Muse Shows What 'Permissions' Really Mean

A security researcher found a way to hijack Meta's Muse AI agent and inherit its access to email, messages, and finances. Meta patched it fast — but the bug is a plain re

Launch Episode: Anthropic's New Dashboard: Measuring the Pace, Not Proving the Brakes

Anthropic has published self-reported metrics on how much of its own AI R&D is done by Claude, how tightly it monitors its 30,000 internal agents, and how much compute go