The Fabric of AI Governance

How Domestic Regulations and International Frameworks Tackle Ten Concerns on AI

Report Image

Our report systematically compares how governments address ten common concerns on AI at the domestic and international level.

Authors

Tommaso Giardini, Nils Deeg

Date Published

16 Nov 2026

As governments worldwide start regulating AI, this report reveals which policy concerns they prioritise and what regulatory requirements they impose to tackle each concern. 

  • We focus on ten concerns: transparency, safety, consumer protection, data governance, human oversight, discrimination and bias, information integrity, labour, intellectual property, and accountability. Governments currently prioritise transparency and safety in AI rulebooks.

  • We distinguish 50 types of regulatory requirements that governments use to address these ten concerns. The most common regulatory requirements include disclosure, information sharing, and documentation (transparency), as well as risk and security assessment (safety).

We deconstruct AI governance at the international and domestic level, on two layers of detail: 

  • To give a broad overview of global activity, we use the Digital Policy Alert (DPA) database, which tracks over 28,000 global digital policy developments. We find 1,756 entries that target the AI sector specifically and address one of the ten concerns.

  • To provide detailed insight on regulatory approaches, we use the full-text libraries on domestic AI regulation and international AI coordination on Clairk. At the domestic level, we analyse 84 AI laws and regulations, finding 454 provisions that establish 935 regulatory requirements. At the international level, we investigate nine international AI frameworks, identifying 203 mentions of the ten concerns. 

The three sections below analyse how governments address each concern in more depth. 

  • The first section briefly summarises our findings on the domestic and international level, showing whether the two levels are aligned, across all ten concerns.

  • The second section synthesises domestic and international activity for each concern.

  • The third section breaks down the language of international AI frameworks on each concern.

Beyond the report, readers can engage with our data in real time with unprecedented sophistication: 

  • The DPA’s AI Governance Hub dedicates a page to each concern, showing both the relevant DPA entries and the provisions of domestic AI regulations and international AI frameworks in full text, structured using our taxonomy. 

  • The DPA’s Activity Tracker provides daily updates on government regulation of the digital economy, beyond AI. Every entry provides a description and the official government source. Users can filter and download the dataset and set up tailored notifications.

  • The St.Gallen Endowment’s technology suite empowers users to interact with our data, including our text analysis tool Clairk, our chat system (drawing exclusively from verified data), and our API and MCP integrations, which feed our data directly into users’ systems.

We thank the Patrick J. McGovern Foundation for the generous support that made this report and the Digital Policy Alert’s AI Governance Hub possible. This report considers policy developments until 22 October 2026 but considers more recent real-life examples. It builds on this methodology.