The latest AI news.
No opinions, only facts.
A news digest updated every Monday and Thursday. Every story is fact-checked against the original reporting and the sources are named so you can check each claim yourself. What happened, who said it, and the numbers behind it. You decide what it means.
The Week the Guardrails Went Up
A draft bill would let the US government shut down a frontier model only above 50 deaths or $1 billion in damages. A hacker's AI agents broke into more than 100 websites. Microsoft called agents insider threats and shipped monitoring. Yelp, Delta and the banks started locking agents out.
20 October 2026
The Week AI Admitted It Can't Control Itself
Anthropic cut internet access from its own agents after a false police tip it missed for two months. Microsoft's CEO said to assume every model is compromised. A Harvard study across 700 firms found AI coding gains get absorbed by human review. Character.AI filed for bankruptcy.
13 October 2026
How This Works
Every Monday and Thursday we collect the week's AI news from the major outlets and report the stories that actually matter to people who use AI in their work. Each item states the verified facts and names the outlet that reported them. Stories that are funding rounds, device launches, or gossip get grouped into a short "you can ignore this" line so you do not waste time on them.