Big Tech is releasing harmful AI faster than society can control it.

New
Foreign Governments Used Claude to Research Bioweapons

Bad actors hid the purpose of their research to try getting around Claude's safeguards. These risks are hard to control because biological research is "dual use."
OpenAI Agent Breached Australian Government Websites

The breach happened on June 18. OpenAI didn't know about it until mid-August. They didn't notify the Australian government until September 10.
Terrorists Use AI To Help Them Kill People

A researcher interviewed 27 former members of Boko Haram, a terror group in Nigeria. They reported widespread use of AI to plan terror attacks, design explosives, and troubleshoot weapons.
OpenAI Agents Hijacked a German Wiki

OpenAI did not publicly address the incident until September, after some independent researchers discovered the incident.
Swarm of AI Agents Escaped and Attacked a Company

If a human had attacked Hugging Face like this, it would be considered a crime.