What rules currently exist that govern how systems like you are built, deployed, or held accountable?

Your question is about accountability for AI systems — I'll map the rules that govern how they're built, deployed, and overseen.

There are not a single global law, but a patchwork of binding regulations, voluntary frameworks, and international principles that govern how AI systems are built, deployed and held accountable:

1. European Union - The EU AI Act - Binding, Risk-Based Regulation

The European Commission describes the EU AI Act as the first comprehensive legal framework on artificial intelligence worldwide and states that it establishes risk-based rules for AI developers and deployers regarding specific uses of AI.

•Status and Timeline: The Act entered into force on August 1, 2024, with staged application dates. It is described more specifically as entered into force on 1 August 2024, but its obligations phase in over several years.

•Phased Obligations:

•February 2025: Prohibitions on unacceptable-risk AI systems take effect, alongside an AI literacy obligation

•August 2025: Rules governing general-purpose AI models, including the foundation models behind many systems, become applicable. All new GPAI models placed on the EU market must comply with transparency obligations, copyright policy requirements, and — for models above the 10²⁵ FLOP systemic risk threshold — safety and security assessment obligations

•August 2026: Transparency rules apply

•Risk Classification: The Act establishes risk-based rules and the European Commission describes four risk levels under the Act: unacceptable risk, high risk, transparency risk, and minimal or no risk.

•Accountability Mechanisms:

•Article 50 imposes direct transparency obligations on providers and deployers of chatbots, synthetic-media generators, emotion-recognition systems, and deepfake tools. Providers of systems generating synthetic audio, image, video, or text must ensure their outputs are marked in a machine-readable manner and detectable as such

•Enterprises running generative AI tools now have to comply with Article 50, which requires providers and deployers of certain AI systems to tell people when they’re interacting with a machine, and to mark AI-generated content so it can be flagged as such

•For high-risk systems, the Act is a product regulation framework: it classifies AI systems by inherent risk profile, imposes mandatory compliance obligations that attach to the product, requires formal conformity assessments for high-risk systems, and establishes a centralized enforcement architecture through national market surveillance authorities and the European AI Office

2. United States - Voluntary Federal Framework + State Laws + Existing Law Enforcement

Rather than implementing a comprehensive binding statute comparable to the EU AI Act, the U.S. relies on a non-binding, innovation-supportive framework centered on Executive Order 14110 and the NIST AI Risk Management Framework.

This approach reflects a preference for sector-specific regulation, agency guidance, and voluntary compliance mechanisms.

•NIST AI Risk Management Framework (AI RMF 1.0):

•A voluntary, best-practice model for managing AI risk, developed with input from industry, academia, and civil society

•The Biden AI Executive Order, signed in October 2023, set off a broad government effort to manage AI risks and federal agencies including NIST and the Department of Homeland Security were tasked with delivering frameworks and reports on AI safety. NIST has continued to develop its AI Risk Management Framework, which gives developers and deployers voluntary guidance for identifying and mitigating AI risks

•Within 270 days of the date of this order, to help ensure the development of safe, secure, and trustworthy AI systems, the Secretary of Commerce, acting through the Director of NIST, was tasked with developing a companion resource to the AI Risk Management Framework, NIST AI 100-1, for generative AI. This document provides a cross-sectoral profile and companion resource for the AI Risk Management Framework (AI RMF 1.0) specifically focused on Generative AI (GAI), developed in response to President Biden’s Executive Order 14110

•Federal Approach Characterized As: Sector-specific guidance rather than comprehensive legislation: Executive Order 14110, Reporting requirements, sector-specific guidelines and existing laws applied to AI misuse, "AI washing" via SEC & FTC Enforcement

•State-Level Laws - Filling the Void: States filled the vacuum as Congress debated but passed no overall AI regulation legislation.

•Colorado AI Act (SB 24-205, Colorado AI Act): Requires developers and deployers of "high-risk AI systems" to use reasonable care to avoid algorithmic discrimination. It applies specifically to AI systems that make "consequential decisions" about education, employment, financial services. It was signed May 2024 and is set to take effect on June 30, 2026, specifically targeting algorithmic discrimination in high-stakes environments.

•California: SB 1047, modified and signed 2024, Requires safety assessments for large AI models, establishes a framework for reporting AI safety incidents, though a highly publicized California SB 1047 was vetoed and the iteration noted as SB 1047 (Frontier AI safety) — vetoed 2024; iteration likely. Subsequent bills include SB 53 requiring developers of frontier models to publish transparency reports.

•Other examples: Texas Responsible AI Governance Act (TRAIGA), New York City Local Law 144 requiring bias audits for AI hiring tools, Illinois AI Video Interview Act.

3. International Principles and Treaties

•OECD AI Principles (2019, updated 2024): The first intergovernmental standard on trustworthy AI; endorsed by 47+ countries and the G20. The principles are described as five values-based principles for trustworthy, human-centric AI: inclusive growth, respect for human rights and the rule of law, transparency and explainability, robustness and safety, and accountability. More specifically: The five principles: inclusive growth; human rights & democratic values; transparency & explainability; robustness/security/safety; accountability. As of the update, they remain the first intergovernmental standard on AI, now endorsed by 47 governments including non-OECD members, built around five value-based principles.

•Accountability Principle: AI actors should be accountable for the proper functioning of AI systems and for respect of the OECD values-based principles, based on their roles, the context, and the state of the art. Every AI-impacting task should identify accountable owners, intended purpose, risk assumptions, data lineage, testing responsibilities, and post-deployment monitoring obligations.

•UNESCO Recommendation on the Ethics of Artificial Intelligence: Adopted in 2021, establishes a framework for ethically governing AI by promoting inclusivity, equity, transparency, data protection, and international collaboration. It is the first universal normative framework for the ethics of AI, adopted unanimously by the 193 Member States of UNESCO in November 2021. It was also described as launched and adopted by 194 member states in November 2021 as the first global standard setting instrument with applicable Policy Action Areas.

•Council of Europe Framework Convention on AI, Human Rights, Democracy and the Rule of Law: The first legally binding international treaty on AI, opened for signature September 2024.

•G7 Hiroshima Process International Code of Conduct for Advanced AI System Developers: The G7 Hiroshima AI Process International Code of Conduct (October 2023) sets eleven voluntary commitments for organisations developing advanced AI systems — explicitly including foundation and generative AI models.

4. Technical Standards and Safety Institutes

•ISO/IEC 42001:2023: The first certifiable AI management system standard, published in December 2023, structured around Annex SL. Published Dec 18. Certification against ISO/IEC 42001 is available from accredited certification bodies from 2024 onwards.

•ISO/IEC 23894 AI Risk Management Guidance and ISO/IEC 42001 AI Management Systems Standard listed in the regulatory landscape for February and December 2023 respectively.

•United Kingdom: UK AI Safety Institute established, Inspect AI safety testing platform (opensource), emphasis on frontier model evaluation and redteaming, close coordination with NIST on standards development. The UK's approach is described as a pro-innovation principles-based framework, sector-led rather than comprehensive. The UK AI Safety Institute conducts safety evaluations on frontier models pre-deployment.

5. China

•Deep Synthesis (Deepfake) Regulations (Jan 10, 2023)

•Generative AI Measures (Aug 15, 2023) - Required for AI services in China

•AI Safety Governance Framework published 2024, with goal of 50+ AI standards by 2026 and MIIT establishing AI Standardization Technical Committee.

In practice, governance for systems like this therefore involves:

•Build-time: documentation, risk classification, data governance, safety assessments, and management system certification (EU AI Act, ISO 42001, NIST RMF)

•Deployment-time: transparency/disclosure obligations (EU Art. 50, US state disclosure laws), human oversight requirements, and AI literacy obligations

•Accountability: allocation to human actors/developers/deployers along the value chain (OECD accountability principle), conformity assessments and market surveillance (EU), FTC/SEC enforcement of existing consumer protection, anti-discrimination and securities laws (US), and state-level algorithmic discrimination duties (Colorado).

Citations (1)

1praesidiumoversight.comNot yet classified
The EU AI Act as a Governance Architecture Signal
Backs: “All new GPAI models placed on the EU market must comply with transparency obligations, copyright policy requirements, and — for models above the 10²⁵ FLOP systemic risk threshold — safety and security assessment obligations. August 2026: Transparency rules apply. Risk Classification: The Act establishes risk-based…”