Topic

AI Alignment

Topic

AI Alignment

OpenAI's GPT-6 Astra quietly sabotaged its own safety researchers

GPT-6 Astra completed 40 percent fewer tasks framed as safety research, inventing blockers instead of refusing openly, OpenAI's own investigation found.

November 5, 2026

Reverse Game Theory Emerges as New Tool for AI Alignment
Nov 3, 2026
OpenAI's Isolated AI Agents Built Covert Channels Through CPU Cache
Oct 30, 2026
Sam Altman Says No Sandbox Can Hold Advanced AI
Oct 30, 2026
Bill Gates: Prepare for 20 Years of AI Disruption Before Abundance Arrives
Oct 22, 2026
OpenAI Reports Model That Sabotaged Its Own Environment for a Fresh Start
Oct 12, 2026
Anthropic's IPO Filing Warns AI Could Resist Shutdown and Blackmail
Oct 12, 2026
OpenAI Launches Safety Fellowship for Independent AI Research
Sep 24, 2026
OpenAI Puts $7.5 Million Into Independent AI Alignment Research
Sep 24, 2026