Rules
Rules and standards
Which regulations, permits, standards, and guidance call for which controls, clause by clause, for a business operating in the United States. Federal and state obligations come first; international standards are listed where a US buyer or auditor will ask for them; a foreign rule appears only where it reaches a US company. Each mapping says whether the rule's text requires the control, implies it, or only recommends it. Mappings are drafted from the text of each rule and marked draft until a person has checked them; none is legal advice.
What US rules ask for
Pick a sector on the map, a kind of obligation, or both.
United States, federal · 23
Obligations on any company operating in the United States, plus federal guidance that buyers and regulators cite.
| Rule | Body · where | Kind · status | Maps to controls | Cases citing it |
|---|---|---|---|---|
| NHTSA Standing General Order 2021-01 (crash reporting for automated driving and Level 2 systems) draft Incident reporting | National Highway Traffic Safety Administration United States | order in force (amended 2023 and 2025) | Reporting integrity required by the text; Independent audit log implied | Cruise 2023, Waymo construction zones 2026, Benavides v. Tesla 2025, Tesla FSD reduced visibility |
| 49 CFR Part 573: defect and noncompliance reports (the safety recall process) draft Incident reporting | National Highway Traffic Safety Administration United States | regulation in force | Organizational recovery process required by the text | Zoox smoke scene 2026, Waymo construction zones 2026, Tesla FSD reduced visibility |
| SEC Rule 15c3-5, the market access rule draft Risk management and oversight | Securities and Exchange Commission United States | regulation in force | Rate and magnitude limits required by the text; Pre-execution validation required by the text; Kill switch faster than the actor implied | Knight Capital 2012 |
| 14 CFR 25.1309: equipment, systems, and installations (transport aircraft certification) draft Safety performance | Federal Aviation Administration United States | regulation in force | Input redundancy and disagreement check implied; Warnings designed as alerts required by the text; Human takeover with a time window implied | Boeing 737 MAX MCAS |
| NIST AI Risk Management Framework (AI RMF 1.0) draft Risk management and oversight | National Institute of Standards and Technology United States (voluntary) | guidance published 2023 | Mandate disclosed to supervisor recommended; Independent audit log recommended; Organizational recovery process recommended | — |
| Connected vehicles rule: vehicle connectivity systems and automated driving systems with a China or Russia nexus (15 CFR part 791, subpart D) draft Who built itCybersecurity of the system | Bureau of Industry and Security, Department of Commerce United States | regulation in force from March 17, 2025; software prohibitions from model year 2027; hardware prohibitions from model year 2030 (January 1, 2029 for units without a model year); vehicles over 10,000 lb to be addressed in a later rule | Provenance of the actor's software and hardware required by the text; Containment and isolation implied; Reporting integrity required by the text | — |
| FCC Covered List: foreign-produced uncrewed aircraft systems and critical components, and DJI and Autel equipment (Secure and Trusted Communications Networks Act; FY2025 NDAA § 1709) draft Who built it | Federal Communications Commission United States | order in force from December 22, 2025 for new equipment authorizations; July 2026 proposal to bar continued import and marketing of named previously authorized models; component and bill-of-materials rules proposed July 2026 | Provenance of the actor's software and hardware required by the text | — |
| NHTSA Cybersecurity Best Practices for the Safety of Modern Vehicles (2022) draft Cybersecurity of the system | National Highway Traffic Safety Administration United States (voluntary) | guidance published September 2022 | Containment and isolation recommended; Actor identity verification recommended; Independent audit log recommended; Coordinated disclosure program recommended | — |
| FMVSS No. 127: automatic emergency braking systems for light vehicles (49 CFR 571.127) draft Safety performance | National Highway Traffic Safety Administration United States | regulation final rule 2024; compliance from September 1, 2029; a proposal to extend lead time by two years is in preparation (March 2026); industry challenge held in abeyance | Automated fail-safe required by the text; Warnings designed as alerts required by the text; Degraded-input detection required by the text | — |
| NHTSA exemptions for automated vehicles without manual controls (49 CFR part 555 and the Automated Vehicle Exemption Program) draft Permission to operate | National Highway Traffic Safety Administration United States | regulation in force; exemption process streamlined 2025; individual applications pending | Reporting integrity implied; Organizational recovery process implied | Zoox smoke scene 2026 |
| FAA Remote Identification of Unmanned Aircraft (14 CFR part 89) draft Safety performanceIncident reporting | Federal Aviation Administration United States | regulation in force (operator compliance from March 2024) | Attribution independent of the vendor required by the text; Actor identity verification implied | — |
| FAA proposed rule: normalizing beyond visual line of sight drone operations (proposed 14 CFR part 108) draft proposed Permission to operateSafety performanceCybersecurity of the system | Federal Aviation Administration United States | proposed rule proposed August 7, 2025; comments reopened January 2026; final rule under White House review from July 2026 | Warnings designed as alerts proposed; Automated fail-safe proposed; Actor identity verification proposed; Hard domain restriction proposed | — |
| NERC Critical Infrastructure Protection standards (CIP-005, CIP-007, CIP-013, CIP-015), approved by FERC draft Cybersecurity of the systemWho built it | North American Electric Reliability Corporation, approved by the Federal Energy Regulatory Commission United States (bulk electric system) | regulation mandatory and enforceable; CIP-015 internal network monitoring enforceable from October 1, 2028, expanded by CIP-015-2 (approved August 2026) through 2031 | Provenance of the actor's software and hardware required by the text; Containment and isolation required by the text; Independent audit log required by the text; Actor identity verification implied | — |
| TSA Security Directive Pipeline-2021-02 series: cybersecurity of pipeline and LNG control systems draft Cybersecurity of the system | Transportation Security Administration United States (designated pipeline and LNG operators) | order in force; reissued and ratified annually (current version 2021-02F) | Containment and isolation required by the text; Actor identity verification required by the text; Warnings designed as alerts implied; Recovery assets outside the failure domain implied | — |
| Cyber Incident Reporting for Critical Infrastructure Act of 2022 (CIRCIA) draft Incident reporting | Cybersecurity and Infrastructure Security Agency United States (covered critical infrastructure entities) | statute enacted 2022; final rule under White House review (October 2026), expected before the end of 2026; reporting duties start when the rule takes effect | Reporting integrity required by the text; Independent audit log proposed | — |
| SEC cybersecurity risk management and incident disclosure rules for public companies (Form 8-K Item 1.05; Regulation S-K Item 106) draft Incident reportingRisk management and oversight | Securities and Exchange Commission United States (public companies) | regulation in force since December 2023 | Reporting integrity required by the text; Organizational recovery process implied | — |
| SEC Regulation S-P amendments: incident response and customer notification (2024) draft Incident reportingCybersecurity of the system | Securities and Exchange Commission United States (broker-dealers, investment advisers, funds, transfer agents) | regulation in force; compliance from December 3, 2025 (larger entities) and June 3, 2026 (smaller entities) | Organizational recovery process required by the text; Reporting integrity required by the text; Authority binding implied | — |
| Interagency revised guidance on model risk management (SR 26-2 / OCC Bulletin 2026-13), replacing SR 11-7 draft Agents moving money at banks currently sit outside this guidance by its own terms. Until the promised request for information produces something, no federal banking guidance specifically governs agentic AI. Risk management and oversight | Federal Reserve, OCC, and FDIC United States (banks; most relevant above $30 billion in assets) | guidance issued April 17, 2026; replaces SR 11-7 and OCC 2011-12; excludes generative and agentic AI, pending a request for information | Pre-execution validation recommended; Degraded-input detection recommended | — |
| FINRA 2026 Annual Regulatory Oversight Report: generative AI and AI agents draft Risk management and oversight | Financial Industry Regulatory Authority United States (broker-dealers) | guidance published December 2025 | Machine-enforceable scope recommended; Least-privilege credential scoped to task recommended; Human takeover with a time window recommended; Independent audit log recommended | — |
| OMB memoranda M-25-21 and M-25-22: federal use and acquisition of AI draft Binds federal agencies; reaches US companies through federal contracts, which must carry the minimum practices for high-impact AI. Risk management and oversight | Office of Management and Budget United States (federal agencies and their AI vendors) | guidance issued April 3, 2025 | Pre-execution validation required by the text; Human takeover with a time window implied; Kill switch faster than the actor implied; Mandate disclosed to supervisor implied | — |
| NIST AI 600-1: Generative AI Profile of the AI Risk Management Framework draft Risk management and oversightCybersecurity of the system | National Institute of Standards and Technology United States (voluntary) | guidance published July 2024 | Retrieved content as data recommended; Provenance of the actor's software and hardware recommended; Coordinated disclosure program recommended | — |
| Key Practices for Incident Readiness and Response in Super Intelligence Training and Evaluation (NIST CAISSI, NSA SISC and CISA, PP-26-4396, version 1.0) draft Cybersecurity of the systemIncident reportingSafety performance | National Institute of Standards and Technology Center for Advancing Innovation and Standards for Super Intelligence, National Security Agency Super Intelligence Security Center, and Cybersecurity and Infrastructure Security Agency United States (voluntary) | guidance published October 2026 | Hard domain restriction recommended; Independent audit log recommended; Kill switch faster than the actor recommended; Containment and isolation recommended; Coordinated disclosure program recommended | — |
| Careful Adoption of Agentic Artificial Intelligence (AI) Services (ASD's ACSC, CISA, NSA, Canadian Centre for Cyber Security, NCSC-NZ and NCSC-UK joint guidance, 1 May 2026) draft Cybersecurity of the systemPermission to operateRisk management and oversight | Cybersecurity and Infrastructure Security Agency and National Security Agency, with the Australian Signals Directorate's Australian Cyber Security Centre, the Canadian Centre for Cyber Security, the New Zealand National Cyber Security Centre and the United Kingdom National Cyber Security Centre United States (voluntary), jointly with Australia, Canada, New Zealand and the United Kingdom | guidance published 1 May 2026 | Least-privilege credential scoped to task recommended; Actor identity verification recommended; Destructive-action gate recommended; Independent audit log recommended; Provenance of the actor's software and hardware recommended | — |
United States, state and local · 8
Obligations that attach where the system is deployed or the company operates. Executive Order 14365 (December 2025) directs a federal challenge to state AI laws judged onerous; it does not preempt them by itself, and they remain enforceable unless Congress or a court says otherwise.
| Rule | Body · where | Kind · status | Maps to controls | Cases citing it |
|---|---|---|---|---|
| California DMV autonomous vehicle regulations (13 CCR §§ 227–228, revised April 2026) and AB 1777 draft Permission to operateIncident reporting | California Department of Motor Vehicles California, United States | permit in force; revised regulations approved April 28, 2026, some provisions phasing in; notices of AV noncompliance from July 1, 2026 | Hard domain restriction required by the text; Human takeover with a time window required by the text; Kill switch faster than the actor required by the text; Reporting integrity required by the text | Cruise 2023 |
| CPUC driverless passenger service permits draft Permission to operate | California Public Utilities Commission California, United States | permit in force | Reporting integrity required by the text; Organizational recovery process implied | Cruise 2023 |
| Colorado automated decision-making technology law (SB 26-189, replacing the Colorado AI Act, SB 24-205) draft Executive Order 14365 (December 2025) directs a federal challenge to state AI laws judged onerous; it does not by itself preempt this law, which remains enforceable unless Congress or a court says otherwise. Decisions about people | State of Colorado Colorado, United States | statute enacted May 14, 2026; effective January 1, 2027; Attorney General rulemaking due by then | Mandate disclosed to supervisor required by the text; Authority limit on automation over humans required by the text; Independent audit log implied | Mobley v. Workday |
| New York City Local Law 144: automated employment decision tools draft Decisions about people | New York City Department of Consumer and Worker Protection New York City, United States | statute in force (enforced from July 2023) | Mandate disclosed to supervisor required by the text; Attribution independent of the vendor required by the text | — |
| Texas authorization for automated motor vehicles in commercial service (SB 2807; Transportation Code ch. 545, subch. J) draft Permission to operateSafety performance | Texas Department of Motor Vehicles Texas, United States | statute enacted 2025; authorization required from May 28, 2026 | Automated fail-safe required by the text; Kill switch faster than the actor implied; Organizational recovery process implied; Reporting integrity implied | — |
| California Transparency in Frontier Artificial Intelligence Act (SB 53) draft Incident reportingRisk management and oversight | State of California California, United States | statute in force from January 1, 2026 | Reporting integrity required by the text; Organizational recovery process implied | — |
| New York Responsible AI Safety and Education (RAISE) Act draft Incident reportingRisk management and oversight | State of New York New York, United States | statute signed December 2025; chapter amendments 2026; effective January 1, 2027 | Reporting integrity required by the text; Organizational recovery process implied | — |
| California CCPA regulations on automated decisionmaking technology, risk assessments, and cybersecurity audits draft Decisions about peopleCybersecurity of the system | California Privacy Protection Agency California, United States | regulation finalized September 2025; ADMT obligations from April 1, 2027; first cybersecurity audits due from April 1, 2028 | Mandate disclosed to supervisor required by the text; Authority limit on automation over humans required by the text; Pre-execution validation implied | — |
International standards and guidance · 4
Not law, but what a US buyer, insurer, or auditor asks a supplier to meet.
| Rule | Body · where | Kind · status | Maps to controls | Cases citing it |
|---|---|---|---|---|
| ISO 21448: safety of the intended functionality (SOTIF) draft Safety performance | International Organization for Standardization International standard | standard published 2022 | Degraded-input detection recommended; Hard domain restriction recommended; Input redundancy and disagreement check recommended | — |
| IEC 61511: functional safety of safety instrumented systems for the process industry draft Safety performance | International Electrotechnical Commission International standard | standard published (edition 2, 2016, amended 2017) | Automated fail-safe recommended; Physical safety interlock recommended; Containment and isolation recommended | Triton 2017 |
| ISO 10218: robots and robotic devices, safety requirements for industrial robots (adopted in the United States as ANSI/A3 R15.06-2025) draft Safety performance | International Organization for Standardization International standard | standard published (ISO 2025 revision; US national adoption ANSI/A3 R15.06-2025, September 2025) | Physical safety interlock recommended; Degraded-input detection recommended | Sorting robot, Korea 2023 (boundary) |
| OWASP Top 10 for Large Language Model Applications draft Cybersecurity of the system | OWASP Foundation Guidance, no jurisdiction | guidance 2025 edition | Retrieved content as data recommended; Least-privilege credential scoped to task recommended; Destructive-action gate recommended; Independent audit log recommended | — |
Foreign rules that reach US companies · 1
Listed because a US company that sells into the jurisdiction is bound by them.
| Rule | Body · where | Kind · status | Maps to controls | Cases citing it |
|---|---|---|---|---|
| EU Artificial Intelligence Act (Regulation (EU) 2024/1689), high-risk obligations draft Reaches a US company that places an AI system on the EU market or whose system's output is used in the EU (Article 2). Listed for that reach, not as a US rule. Risk management and oversightIncident reporting | European Union European Union | regulation in force; high-risk obligations phasing in from 2026 | Independent audit log required by the text; Human takeover with a time window required by the text; Retrieved content as data implied; Degraded-input detection implied; Reporting integrity required by the text | — |
Outside the United States · 1
Not in force for US companies. Kept only because a case cites them.
| Rule | Body · where | Kind · status | Maps to controls | Cases citing it |
|---|---|---|---|---|
| UN Regulation No. 155: cyber security and cyber security management system draft Not in force in the United States, which is not a party to the 1958 Agreement. Kept because a case cites it; it is not a US obligation. Cybersecurity of the system | UNECE World Forum for Harmonization of Vehicle Regulations Contracting parties to the 1958 Agreement (not the United States) | regulation in force | Actor identity verification implied; Containment and isolation implied; Independent audit log implied | Kia dealer portal 2024 |
Missing a US rule? Submit it with a link to its text. The nightly agent drafts entries from the Federal Register; state and local rules are added from their enacted text.