This analysis describes what Anthropic's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
The clause identifies CBRN misuse as the primary threat model driving ASL-3 deployment safeguard design, establishing the scope of harm Anthropic is specifically targeting at that capability level.
Anthropic's ASL-3 deployment safeguards are being built to restrict AI capabilities that could be misused for severe CBRN-related harm.
How other platforms handle this
your parent(s) or legal guardian(s) are responsible for: (i) your activities while using our Services; (ii) your compliance with this Policy; and (iii) ensuring your use of the Services doesn't violate any child protection laws.
You are not permitted to access or use our Services in any jurisdiction or country if it would be contrary to the law or regulation of that jurisdiction or if it would subject us to the laws of...
Don't create or share defamatory or libelous content that could damage someone's reputation or falsely portray them in a harmful way.
"These safeguards are being designed to prevent misuse of capabilities that could enable severe harm, particularly in relation to chemical, biological, radiological, and nuclear (CBRN) technologies.Excerpt from Anthropic's Responsible Scaling Policy
Get the weekly coverage letter
Companies change their terms without announcing it. We capture governance documents, compare each new version with the last, and publish what changed, with the source text behind it. One email a week on what we found and what we could not reach.
The clause identifies CBRN misuse as the primary threat model driving ASL-3 deployment safeguard design, establishing the scope of harm Anthropic is specifically targeting at that capability level.
Anthropic's ASL-3 deployment safeguards are being built to restrict AI capabilities that could be misused for severe CBRN-related harm.
ConductAtlas has identified this type of provision across 238 platforms. See the full comparison.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Anthropic.