The Financial Times, citing investor prospectus filings with the SEC and an independent review by research group METR, reported that Anthropic has activated stricter safety safeguards due to rapid acceleration in its internal model development. According to METR’s assessment, internal deployment of AI tools has accelerated capability research progress by roughly 1.9x over baseline rates. Crossing this threshold under Anthropic’s Responsible Scaling Policy triggers mandatory security enhancements to govern AI-assisted R&D workflows.
Tags: #Anthropic #METR #AISafety #Research #ResponsibleScaling
Sources: Financial Times | METR | SEC Filings
R&D Acceleration Threshold: Anthropic Tightens Safety Controls Following METR Assessment