“we will also develop a safety case, which is an assessable argument showing how severe risks associated with a model's CCLs have been minimised to an acceptable level.”
This analysis describes what Google DeepMind'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 commitment to develop a safety case that is 'assessable' and must show minimisation to an 'acceptable level' creates a defined, evaluable standard rather than a discretionary or informal review.
Readers are informed that before deployment, Google DeepMind is committed to producing a safety case that must demonstrate—in an assessable way—that severe risks from a model's critical capability levels have been minimised to an acceptable level.
How other platforms handle this
Investing in industry-leading approaches to advance safety and security research and benchmarks, pioneering technical solutions to address risks, and sharing our learnings with the ecosystem.
Our security procedures mean that we may ask to verify your identity before we disclose personal information to you.
By providing us with information about your contacts you certify that you have permission to provide that information to PayPal for the purposes described in this Privacy Statement.
"we will also develop a safety case, which is an assessable argument showing how severe risks associated with a model's CCLs have been minimised to an acceptable level.Excerpt from Google DeepMind's Frontier Safety Framework
Get the research letter
Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean.
The commitment to develop a safety case that is 'assessable' and must show minimisation to an 'acceptable level' creates a defined, evaluable standard rather than a discretionary or informal review.
Readers are informed that before deployment, Google DeepMind is committed to producing a safety case that must demonstrate—in an assessable way—that severe risks from a model's critical capability levels have been minimised to an acceptable level.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Google DeepMind.