FilingsGPT · Safety Frameworks
Amazon · Frontier Model Safety Framework · v2.7 · 2025
How it maps to Article 44-B
Amazon's Frontier Model Safety Framework sets Critical Capability Thresholds and emphasizes quantifying uplift — whether a model materially increases an actor's ability to conduct CBRN, offensive-cyber, or autonomous AI-R&D attacks. A model that meets a threshold is not deployed unless mitigations are shown to be effective, with pre-deployment and safeguards evaluations reviewed by senior leadership and integrated with AWS's enterprise security controls. Its uplift-based thresholds tied to release decisions illustrate Article 44-B's duties to assess catastrophic risk, evaluate models before deployment, and govern how risk is managed.
44-B obligations mapped
An illustrative reference entry — Amazon’s own voluntarily-published safety framework, shown for the kind of disclosure Article 44-B contemplates. It is not an official Article 44-B safety & security protocol (the regime takes effect January 1, 2027). Verify the version, date, and any specific figure against the live source.
FilingsGPT is an independent reference implementation of New York's Article 44-B (the RAISE Act). It is not affiliated with or operated by the NY Department of Financial Services or the State of New York, and is not an official filing system.