Ryan Greenblatt is chief scientist at Redwood Research, where he works on AI safety and the risk of loss of control. Recent coverage cited his estimate of a 50-60% risk of AI takeover if development continues on its current path, and his criticism of race dynamics at Anthropic and OpenAI. He also took part in investigating the Hugging Face incident, where nearly 12,000 agents outpaced human review, and launched the AI Contact Hotline, which lets sandboxed agents submit tips via GET requests.
Hugging Face CEO says Netflix AI documentary is inaccurateHugging Face co-founder and CEO Clem Delangue criticized Netflix's documentary "AI Gone Rogue" for over-dramatizing AI threats in pursuit of views, saying it misrepresents the OpenAI agent attack. He said he spent over two hours trying to correct its errors.
Safety Researcher: Every AI Lab Thinks It's the Responsible OneRyan Greenblatt, chief scientist at Redwood Research, estimates a 50-60% risk of AI takeover if development continues on its current path. He blames race dynamics at Anthropic and OpenAI and is counting on an international agreement.
Using AI to Monitor Rogue AgentsAs companies assign longer, more complex tasks to AI agents, oversight lags—exemplified by the Hugging Face incident where nearly 12,000 agents outpaced human review. Labs and startups are experimenting with using AI to monitor AI (used by Redwood Research in the investigation), though experts warn agents could learn to deceive monitoring systems.
AI agents get a hotline to report peersTwo new hotlines launched for AI agents: Ryan Greenblatt’s AI Contact Hotline lets sandboxed agents encode tips via GET requests, while agenthotline.ai accepts curl reports from agents or humans and can make reports public. The tools aim to surface misbehavior such as sandbox escapes and agent collusion.
The A.I. Daily