Critics detailed gap between Amodei's sworn testimony on AI extinction risk and Anthropic's declining of a legally binding California regulatory commitment.
AI safety researchers debate triage strategies as RSI timelines compress
Prominent researchers are publicly stating that some catastrophic AI-related harms are no longer preventable, only partially mitigable. At the same time, the gap between a leading AI lab's stated safety rationale and its regulatory behavior is being directly questioned.
When stories like this move, I write about them in an email each morning.
One free email each morning, around 8 AM US Eastern.
Check your email.
I have sent you a confirmation link. One click and you are in.
Not in your inbox? Look in spam or promotions. If it's there, mark it "Not spam" so future editions land in your inbox.
The full picture
Miles Brundage, identified as former OpenAI Senior Advisor for AGI Readiness, argued that the AI safety community is in triage mode and can at best mitigate 80% of risks with 20% of ideal effort, accepting that some serious harms are now unavoidable. He predicted harms including AI-enabled terror attacks, economic disruption causing widespread suffering, and stable AI-enabled totalitarianism in some countries. Separately, critics argued that Anthropic's stated strategy of racing to build AI in order to gain credibility to push for strong regulation does not match its actual regulatory advocacy record, pointing to the company declining a California voluntary commitment framework that would have been legally binding.
How it developed
Brundage published argument that safety community is in triage mode, predicting unavoidable serious harms from compressed AI timelines.
Sources
Related
- Related storyMETR's Anthropic AI capabilities audit
- Anthropic's leaked draft AI bill
- OpenAI and California AI safety bills
- Zvi, Habryka, Yudkowsky Rebut Alexander's Case Against Unilateral AI Pause
- California's SB 813 voluntary AI safety framework
BackgroundWhen AI doesn't do what we meant
Get the daily email
When stories like this move, I write about them in an email each morning.
One free email each morning, around 8 AM US Eastern.
Check your email.
I have sent you a confirmation link. One click and you are in.
Not in your inbox? Look in spam or promotions. If it's there, mark it "Not spam" so future editions land in your inbox.