Advancing safe, resilient, and accountable AI across the systems society depends on.
The Center for Trustworthy AI in Critical Infrastructure is a nonpartisan public-interest center of The Trustworthy AI Initiative. Its mission is to help the operators, regulators, and public overseers of America's critical infrastructure sectors adopt artificial intelligence in ways that are safe, resilient, transparent, and worthy of public trust. The Center does not replace sector regulators, ISACs, or federal cybersecurity authorities — it complements them with independent, cross-sector guidance focused specifically on AI.
Nonpartisan · Public-interest · Complements CISA, NIST, and sector regulators
Why now
AI is being embedded into the grid, into hospitals, into water systems, into transit — often faster than the governance, oversight, and resilience practices around it. When these systems fail, the consequences are borne by patients, ratepayers, and communities, not vendors. Trustworthy AI in critical infrastructure is now a public-interest priority.
Vision
A decade from now, every operator of a critical service — every utility, hospital, transit authority, water system, and financial institution — will be able to demonstrate that the AI inside their operations is governed, resilient, and accountable to the public it serves.
Flagship program
The Observatory is the Center's continuous research and intelligence program — producing the annual State of AI in Critical Infrastructure Report, quarterly sector outlooks, AI incident analyses, a regulatory tracker, executive briefings, and sector benchmarks. It is the knowledge engine behind the Center's research, publications, policy recommendations, and long-term funding strategy.
The Center organizes its work around the 16 CISA-designated Critical Infrastructure Sectors so it aligns with how Sector Risk Management Agencies already operate.
Where the Center will publish sector deep-dives, readiness assessments, and the first AI Resilience Score™ rubrics.
Clinical decision support, triage, diagnostics, and hospital operations.
Grid operations, demand forecasting, and AI in generation and distribution.
Treatment optimization, leak detection, and operational technology safety.
Sector work opens as founding sector advisors are seated.
Aviation, rail, transit, maritime, and highway AI applications.
Underwriting, fraud detection, and market-integrity AI systems.
Network operations, routing, and AI-driven service management.
Dispatch, response optimization, and AI in public safety operations.
Covered as the Center's advisor network, funding, and cleared-partner arrangements expand.
Foundational IT and cloud services underpinning other sectors.
Supply-chain optimization and precision agriculture.
Process control and safety-critical AI in chemical operations.
AI in industrial control, quality, and supply chain.
Building systems, physical security, and occupant safety AI.
Structural monitoring, flow control, and safety-critical automation.
Responsible AI across the defense supply chain (via cleared advisors).
AI in public buildings, benefits systems, and public-service operations.
Safety-critical AI oversight in the nuclear sector.
Sector framing follows the 16 CISA-designated Critical Infrastructure Sectors (Presidential Policy Directive 21) so the Center's work aligns with how Sector Risk Management Agencies already organize.
A focused set of programs — starting with the Center's signature AI Resilience Score™ — rather than a broad catalog of services.
The Center's signature program. A public, sector-adapted score that measures whether an AI system deployed in critical operations is governed, monitored, and resilient enough to be trusted with essential services.
Structured self-assessments that help utilities, hospitals, and transit agencies understand where their AI use sits relative to sector norms and public expectations.
A crosswalk that operationalizes existing frameworks for infrastructure operators — so a water utility or health system can see, in one place, what NIST AI RMF, ISO/IEC 42001, ISA/IEC 62443, and CISA CPGs actually require of their AI programs.
A public record of AI use, incidents, and lessons learned across critical infrastructure sectors — so the field can learn from failures without waiting for the next crisis.
The Center exists to protect the essential services people depend on — not to promote any specific vendor, technology, or legislation.
Helps operators catch AI failure modes before they reach patients, ratepayers, and riders.
Gives communities a nonpartisan source they can consult about the AI inside services they depend on.
Focuses on AI resilience, not just AI performance — because critical systems have to keep working.
Helps operators translate frameworks into controls that actually reduce downside.
Sets a shared bar so operators can adopt AI confidently, and vendors can compete on trust.
Cross-sector visibility so lessons in one sector strengthen the others.
Regulators, CISA, sector ISACs, and NIST already do essential work. The Center adds the missing piece: an independent, nonpartisan, public-interest voice focused specifically on AI trust and resilience across sectors.
Set cross-sector cyber baselines and coordinate sector risk. The Center complements them with independent, AI-specific guidance.
Publishes the AI RMF and related frameworks. The Center operationalizes them for infrastructure operators.
Share threat intelligence within sectors. The Center adds cross-sector, public-interest AI trust guidance.
Set and enforce sector rules. The Center supports voluntary adoption of trustworthy-AI practice above the regulatory floor.
Get Involved
The Center is in formation. We are seating founding sector advisors — practitioner leaders from healthcare, energy, and water — before publishing sector work. Use this form to share your interest or nominate someone.
Nonpartisan. Public-interest only.
Independent. Complements CISA, NIST, and sector regulators.
Discreet. All submissions reviewed privately.
Or email infrastructure@ethosairegistry.org.