Resources: Foundations of AI Governance

Exercises: Foundations of AI Governance

Resources: Proposals to Govern AI

Exercises: Proposals to Govern AI

Next steps: Programs

Resources: The AI Policy Toolkit

Exercises: The AI Policy Toolkit

Resources: How AI is Governed Today

Exercises: How AI is Governed Today

Resources: Governance at Frontier AI Companies

Exercises: Governance at Frontier AI Companies

In this unit, we'll examine how governments are actually implementing AI governance. We'll focus on three major players - the US, EU, and China - analyzing how their different approaches reflect their values, priorities, and political systems.
Government approaches to AI governance typically fall into distinct regulatory categories, from risk assessment frameworks to disclosure requirements. These categories represent different tools in the policy toolkit, each with their own strengths and limitations. By comparing how different jurisdictions use these tools, we can better understand both technical and political constraints on governance.
Remember that government policies reflect more than just technical considerations - they embody specific political priorities and economic interests. As you analyze these approaches, consider how different political systems and national priorities shape policy choices.

By the end of the unit, you should be able to:
\- Understand the structure of AI policies and regulations worldwide, focusing on the US, EU, and China.
\- Define and explain the purpose of eight regulatory categories of AI policies.


_Spend 15-20 minutes on this exercise_
Choose one of the eight policy levers below:
- Evaluations and risk assessments
- Model registries
- Incident reporting
- Open and closed-source regulations
- Frontier model cybersecurity
- Discrimination requirements
- AI disclosures
- CBRN hazards regulations
For your chosen lever, evaluate:
- What are its goals?
- Who is implementing it?
- How successful has that actor been at achieving its goals thus far?
- Does this action seem likely to achieve its goals in the next five years? Why or why not?


AI Governance Fast-Track

Exercises: How AI is Governed Today

Exercises