Resources: Foundations of AI Governance

Exercises: Foundations of AI Governance

Resources: The AI Policy Toolkit

Exercises: The AI Policy Toolkit

Resources: How AI Systems Work

Exercises: How AI Systems Work

Resources: Proposals to Govern AI

Exercises: Proposals to Govern AI

Resources: The Promise and Perils of AI

Exercises: The Promise and Perils of AI

Resources: Contributing to AI Governance

Exercises: Contributing to AI Governance

Resources: How AI is Governed Today

Exercises: How AI is Governed Today

Resources: Governance at Frontier AI Companies

Exercises: Governance at Frontier AI Companies

Following our units on what governance looks like today – both from governments and from frontier labs – we'll explore concrete proposals for how we might better govern AI in the future. We'll examine four distinct approaches: proactive commitments that define responses to specific scenarios, compute-focused policies that leverage a key technical bottleneck, defensive acceleration strategies that aim to build protective capabilities, and international institution-building to coordinate global efforts to create safe AI systems.

These proposals represent different philosophical approaches to governance. Some focus on defining clear triggers for action, others on controlling key resources, others on building protective capabilities, and others on institution-building. Each makes different assumptions about what's technically feasible, politically possible, and ethically desirable.

As you evaluate these proposals, consider the practical challenges of implementation. What incentives would different actors have to adopt or resist these measures? How might they interact with existing governance structures? What unintended consequences might they create?

By the end of the unit, you should be able to:
\- Analyze specific policy proposals for AI governance and their underlying assumptions
\- Evaluate the tradeoffs, feasibility, and potential effectiveness of different governance approaches
\- Explain how different proposals might complement or conflict with each other


_Spend 20-30 minutes on this exercise_
Imagine it’s 2030. BigLab claims they’ve achieved AGI. You’re a policymaker, and it’s your job to steer it. In this exercise, you’ll create a policy proposal.
First, re-read “Scenario 4: Advanced AI on a Knife Edge” from Session 2’s [“Future Risks of Advanced AI”](https://assets.publishing.service.gov.uk/media/653bc393d10f3500139a6ac5/future-risks-of-frontier-ai-annex-a.pdf#page=21) (page 21). You'll create your policy for this scenario. 
Then, decide what country you’d like to create a policy for. It can be your country or a country you understand well. In 2-3 sentences, describe the world you now live in – what this model does and how it's affecting your country. 
From there, in one sentence, describe the goal you will try to achieve with your policy.
Finally, in 5-6 sentences, propose a policy action designed to achieve your goal: Consider the questions below as you design your policy:
- What am I trying to achieve with this policy?
- What strategic priority does this policy advance?
- What stages of the AI lifecycle does my policy affect?
- Is my policy likely to achieve its goal?
- What obstacles might prevent my policy from working as intended?


Create a policy proposal

AI Governance (2024)

Examine concrete proposals for future AI governance, from compute governance to defensive acceleration.

Proposals to Govern AI

Learn the basics of AI: what it is, how it's trained, and what it's capable of.

How AI Systems Work

Understand the potential benefits and risks of advanced AI systems

The Promise and Perils of AI

Learn why AI governance matters, who the players are, and what technical factors shape our choices.

Foundations of AI Governance

Discover the concrete tools governments can use to steer AI and the hard tradeoffs they face.

The AI Policy Toolkit

Explore how the US, EU, and China are actually governing AI through eight types of regulation.