NIST aims for summer release of AI cyber guidelines
dailyRegulation
Summary
NIST's Center for AI Standards and Innovation (CAISI) plans to publish an AI-specific cybersecurity framework profile “sometime this summer,” issuing sequential draft control overlays for different AI system types: a predictive-AI overlay expected this summer and an agentic overlay in late summer to early fall. CAISI is coordinating with industry and—per a now-removed partners page—had arrangements for Google DeepMind, Microsoft, and xAI to submit models for testing; NIST aims to finalize all guidance by 2027.
Why it matters
Near-term, model-type-specific cybersecurity baselines and firm timelines will directly affect vendor risk, procurement, testing, and compliance planning.