Resources: AI standards and regulations

Exercises: AI standards and regulations

Resources: Introduction to Machine Learning

Exercises: Introduction to Machine Learning

Resources: Career Advice (AIG)

Exercises: Career Advice (AIG)

Resources: Closing regulatory gaps through international agreements

Exercises: Closing regulatory gaps through international agreements

Resources: Projects  (AIG)

Exercises: Projects  (AIG)

Next steps: Programs

Resources: Challenges in achieving AI safety

Exercises: Challenges in achieving AI safety

Resources: Additional proposals

Exercises: Additional proposals

Resources: Introduction to potential catastrophic risks from AI

Exercises: Introduction to potential catastrophic risks from AI

Resources: Closing regulatory gaps through non-proliferation

Exercises: Closing regulatory gaps through non-proliferation

So far, we have studied one cluster of governance ideas: regulating industrial-scale AI development, while reducing risks from states that do not enact regulations. This week, we’ll study additional (often complementary) governance proposals, still with a focus on navigating the impacts of increasingly advanced AI. These additional proposals involve topics including AI companies’ policies, “CERN for AI,” slowing down AI development, and developing positive visions for AI.

By the end of the unit, you should be able to:
\- Explain several strategy ideas for addressing AI safety challenges, and compare their strengths and weaknesses
\- Describe several ways in which AI companies can be important governance actors (i.e. decisions they make that shape the impacts of AI)
\- Identify two reasons why shared positive visions for AI might be very beneficial, and a few challenges with developing such visions  


Read[ our guidance](https://docs.google.com/document/d/e/2PACX-1vTPYrsCZV95S03L5-cO-p10cHz1TBMxSr_onylcSdO7KHNVmBSeOGVTq1LDlO49tXHoAJ_oZh6D4pt1/pub) on how to approach writing to learn exercises.
This week we offered a brief overview of a range of other approaches to addressing extreme risks from AI with AI governance. In this exercise, we encourage you to pick one strategy from this week’s resources that you think sounded particularly promising, and do more research into it.

**Prompt: pick a strategy from this week’s resources and make a case as to why or why not you would prioritise putting it into practice in the next 5 years.** Consider both the impact and feasibility of the strategy being successful.
**Aim to write <= 400 words.**
The goal of this exercise is to deepen your understanding of one strategy of your choice, as well as practise the skill of strategic research. You will try to form a conclusion on whether the strategy is feasible or not, but in the process you’ll likely unearth a range of further questions you’d like to answer to get more detail on its feasibility and strengths.
Examples of strategies from this week's resources (non-exhaustive):
- Supporting and lobbying AI companies to adopt good policies for self-governance
- “CERN for AI"
- Slowing down AI development (various proposals for how to do that listed in the resources)
- Create means for rapid shutdown of large compute clusters and training runs (and other ideas from the ‘12 tentative ideas’ resource)
You may also wish to consider:
- What challenges might it face in being put into practice?
- What are the strengths of this strategy?
- What are the downsides of the strategy, and how would you mitigate those factors?
- To comment on feasibility - how, concretely, would you go about putting it into practice? You may need to make some assumptions to answer this question.


AI Governance (2023)

Exercises: Additional proposals

Exercises