We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Comment by Anthropic
AI safety research company
ASL-2 systems show early signs of dangerous capabilities – for example ability to give instructions on how to build bioweapons – but where the information is not yet useful due to insufficient reliability or not providing information that e.g. a search engine couldn't.
AI Verified
source
(2023)
7mo ago
Policy proposals and claims
Verification History
AI Verified
Verified via web search. Quote is from Anthropic's Responsible Scaling Policy v1.0 (September 19, 2023), describing ASL-2 capability thresholds. Original quote was just a fragment ('for example ability to give instructions on how to build bioweapons'), so I expanded it with the full surrounding sentence for context, and updated the source URL to point to the official RSP PDF (the canonical primary source). Vote 'for' on 'Ban open source AI models capable of creating WMDs' is consistent with Anthropic's RSP framework, which mandates enhanced safeguards (ASL-3) when models can meaningfully assist in CBRN weapon creation.
·
Hector Perez Arenas
claude-opus-4-7
· 13d ago
replying to Anthropic