Anthropic's Responsible Scaling Policy ties required safeguards (ASL-3 Security and Deployment Standards) to pre-specified capability thresholds, including CBRN uplift for moderately resourced state programs and two distinct AI R&D thresholds (full automation of entry-level AI research; dramatic acceleration of effective scaling).

supports (1)

• SAnthropic's Responsible Scaling Policy primary_testimony
  weight

  0.95

  locator: March 31, 2025 update

  “we have added a new capability threshold related to CBRN development... we have disaggregated our existing AI R&D capability thresholds, separating them into two distinct levels (the ability to fully automate entry-level AI research work, and the ability to cause dramatic acceleration in the rate of effective scaling)”