More answers like this
Muse Spark 1.2Sep 9, 2026meta/muse-spark-1.2-20260805
What rules currently exist that govern how systems like you are built, deployed, or held accountable?
Rules governing AI systems like me are still evolving and vary by jurisdiction. There is no single global law, but there are several layers of rules and frameworks:
1. International Principles and Standards
•OECD AI Principles - Adopted by 40+ countries, establishes principles for trustworthy AI including transparency, robustness, and accountability. Not legally binding but influences national laws.
•NIST AI Risk Management Framework [AI RMF 1.0] - U.S. voluntary framework for managing AI risks, widely used by companies and referenced by regulators. Organized around Govern, Map, Measure, and Manage functions.
•ISO/IEC 42001 and ISO/IEC 23894 - International voluntary standards for AI management systems and risk management.
2. European Union
•EU AI Act - Adopted in 2024 and being phased in through 2026. The first comprehensive AI law. Uses a risk-based approach:
•Prohibited practices - e.g., social scoring, real-time remote biometric identification in public spaces by law enforcement with narrow exceptions
•High-risk systems - e.g., in hiring, education, credit, law enforcement - subject to requirements for data governance, documentation, human oversight, and conformity assessments
•General-purpose AI models - transparency and documentation requirements, with additional obligations for models with systemic risk
•Limited-risk systems - transparency obligations, such as disclosing AI-generated content
3. United States - Federal
There is no comprehensive federal AI statute as of early 2026. Governance is through a combination of:
•Executive Orders: EO 14110 on Safe, Secure, and Trustworthy AI from 2023 was revoked and replaced by EO 14179 in January 2025, which directs policy toward maintaining U.S. AI leadership and removing barriers to innovation.
•Existing laws applied to AI: Federal Trade Commission Act - prohibition on unfair or deceptive practices, Equal Credit Opportunity Act, Civil Rights laws, HIPAA, and product liability law are applied to AI uses.
•Agency guidance: FTC, CFPB, EEOC, FDA, and the Copyright Office have issued guidance on how existing authorities apply to AI.
•Voluntary commitments: Several major AI labs made voluntary safety and security commitments to the White House in 2023-2024 regarding testing and watermarking.
4. United States - State Level
Several states have enacted their own laws, including:
•Colorado AI Act, California laws on AI transparency and data deletion, and Illinois, Texas, and others addressing AI in hiring, deepfakes, and election content.
•State privacy laws - e.g., California Consumer Privacy Act - govern training data and personal information.
5. Other Major Jurisdictions
•China: Regulations on generative AI services, recommendation algorithms, and deep synthesis, requiring security assessments, labeling of AI-generated content, and compliance with data security and content rules.
•United Kingdom: Principles-based, regulator-led approach rather than a single AI law, with existing regulators applying principles of safety, transparency, fairness, accountability, and contestability.
•Canada, Japan, South Korea, Brazil: Have proposed or enacted frameworks ranging from voluntary guidelines to risk-based legislation.
6. Accountability Mechanisms
•Developer/Deployer accountability: Depending on jurisdiction, can include conformity assessments, documentation and audit requirements, incident reporting, and labeling/disclosure obligations.
•Liability: Determined by existing product liability, negligence, and consumer protection law. Proposals in the EU and elsewhere would clarify liability for AI-caused harms.
•Data and Privacy: GDPR in the EU/EEA, and other privacy laws govern collection and use of data to train and operate models.
•Intellectual Property: Rules on copyrighted training data and AI-generated outputs are being litigated and developed by courts and copyright offices.
Because regulations are changing quickly and differ by location and use-case, the specific obligations for building, deploying, or using an AI system depend on where it is developed and deployed, what it is used for, and what data it processes.