Draft talk:AI governance

Production disclosure

This article was produced under a structured human-governed workflow using the HAIA-RECCLIN methodology and Checkpoint-Based Governance (CBG). All content was researched, reviewed, and arbitrated by the named human author (Basil C. Puglisi, MPA) at every checkpoint before publication. AI platforms were used as research and editorial tools under explicit human direction and oversight. This workflow is documented at https://github.com/basilpuglisi/HAIA. Attribution: #AIassisted, not AI-generated. BasilPuglisi (talk) 04:59, 28 March 2026 (UTC)

Response to decline — merge target does not exist

@Robert McClenon: Thank you for the review. The decline directs merging content into Regulation of artificial intelligence:, but that page does not exist — the title with the trailing colon is a redlink, and the corresponding Talk page confirms no article exists at that location.

If the intended merge target is Regulation of artificial intelligence (without the colon), I would note that this draft addresses AI governance as a distinct conceptual domain. It covers the ethical/responsible/governance taxonomy (Batool et al. 2025), implementation gap findings (Morley et al. 2020), and international standards frameworks (ISO/IEC 42001, ISO 37000, ISO/IEC 38507), which fall outside the regulatory and legal scope of that article.

Could you clarify the intended target so I can proceed appropriately? BasilPuglisi (talk) 12:38, 31 March 2026 (UTC)

User:BasilPuglisi - Well, you were correct as to the typographical error, and I have corrected it. It was correct in the AFC comment, but asking what was really intended was reasonable. If you think that AI governance should be a separate article, I suggest that you discuss at the talk page of the existing article, Talk:Regulation of artificial intelligence. Robert McClenon (talk) 20:17, 31 March 2026 (UTC)

Case for standalone article: AI governance as a distinct topic

The existing article Regulation of artificial intelligence is scoped to laws, executive orders, national strategies, and enforcement bodies. Its table of contents confirms this: Background, Perspectives, Global guidance, Regional and national regulation (with subsections for over a dozen countries), and Regulation of fully autonomous weapons. The article's own lead sentence defines its scope as "the development of public sector policies and laws for promoting and regulating artificial intelligence."

This draft addresses a distinct topic. The peer-reviewed literature treats AI governance and AI regulation as separate domains:

• Batool, Zowghi, and Bano (2025), in a systematic literature review published in AI and Ethics, defined AI governance as encompassing "who is accountable, what is governed, when governance occurs, and how it is implemented through frameworks, tools, or models." They described regulation as one component within governance, not a synonym for it. They further distinguished ethical AI (moral principles and value alignment), responsible AI (risk minimisation and development safeguards), and AI governance (regulations, procedures, and institutional arrangements). (doi:10.1007/s43681-024-00653-w)

• Stix (2022) documented that terms including "trustworthy AI," "responsible AI," and "ethical AI" have shifted in meaning over time and are often conflated, supporting the need for an article that establishes these distinctions. (doi:10.1007/s44163-022-00041-5)

• Floridi and Cowls (2019) synthesised AI ethics into five principles (beneficence, non-maleficence, autonomy, justice, explicability), a conceptual framework that belongs in a governance article, not a regulation article. (doi:10.1162/99608f92.8cd550d1)

• Morley et al. (2020) reviewed over 100 tools for translating AI ethics into practice and found most were either too abstract or too inflexible for use in development settings. This implementation gap is a governance finding, not a regulatory one. (doi:10.1007/s11948-019-00165-5)

The draft also covers content that has no natural home in the regulation article:

• Standards frameworks: ISO/IEC 42001:2023 (AI management systems), ISO 37000:2021 (governance of organisations), and ISO/IEC 38507:2022 (governance implications of AI use) are governance standards, not regulations.
• Risk management: The NIST AI RMF 1.0 and its GOVERN function address organisational accountability and culture for AI oversight, distinct from statutory requirements.
• Human oversight models: The four oversight models defined in the EU AI Act (human-in-command, human-in-the-loop, human-on-the-loop, human-over-the-loop) are governance design patterns that warrant explanation beyond what a regulation article would provide.
• Geographic and cultural analysis: Jobin, Ienca, and Vayena (2019) found 88% of AI ethics guidelines originated in Europe and North America. Goffi (2022) argued that governance discourse embeds Western-centric assumptions. These are critiques of the governance field itself, not of regulation.
• Implementation evidence: Batool et al. found that only 3 of 28 studies in their review addressed all four governance questions. T

BasilPuglisi (talk) 12:43, 31 March 2026 (UTC)