responsible scaling policy (RSP)
A corporate safety framework that defines risk categories for AI models and commits to running safety evaluations before advancing between categories, with accompanying safeguards required at each level
As defined in AI Timelines and Human Psychology (with Sarah Hastings-Woodhouse)
Other definitions
Anthropic's public commitment framework specifying safety categories for models and thresholds for conducting evaluations before training increasingly powerful systems
AI Timelines and Human Psychology (with Sarah Hastings-Woodhouse)
Where it comes up · 1
Most important claims · 4 of 4Search all claims for “responsible scaling policy (RSP)” →
- Given the vagueness of RSPs and companies' incentive to corner-cut under competitive pressure, there are many technically-compliant ways for companies to continue unsafe development while nominally adhering to their own policies.AI Timelines and Human Psychology (with Sarah Hastings-Woodhouse)
- OpenAI's Preparedness Framework is more honest than Anthropic's Responsible Scaling Policy because it explicitly calls itself a 'living document' and acknowledges that the science of evaluating advanced models is immature and will change, whereas Anthropic's RSP uses the language of binding 'public commitment' despite the same epistemic limitations.AI Timelines and Human Psychology (with Sarah Hastings-Woodhouse)
- Current AI safety plans (Anthropic's Responsible Scaling Policy, OpenAI's Preparedness Framework, DeepMind's policies) are vague commitments rather than concrete plans, as they do not specify which evaluations will be run, at what frequency, with which compute thresholds, or what success criteria trigger pause decisions.AI Timelines and Human Psychology (with Sarah Hastings-Woodhouse)
- AI safety plans from major labs (Anthropic, OpenAI, Google DeepMind) are not actually concrete plans but rather vague policy commitments that fail to specify which evaluations will be run, how often they will be run, or what specific actions trigger pauses in development.AI Timelines and Human Psychology (with Sarah Hastings-Woodhouse)