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

Next steps: Programs

Resources: How AI is Governed Today

Exercises: How AI is Governed Today

Resources: Governance at Frontier AI Companies

Exercises: Governance at Frontier AI Companies

Welcome to the course! This unit focuses on the foundational concepts in AI that you'll need to know to understand the rest of the course.

We'll learn the basics of AI: what it is, how it works, and what it can do right now. We’ll examine the crucial architecture behind deep learning (the leading approach to machine learning): neural networks. We'll also do a deep dive into large language models (LLMs), the technology behind most state-of-the-art foundation models like ChatGPT.

By the end of the unit, you should be able to:
\- Understand that ML systems are good at predicting outputs based on a set of inputs
\- Explain to a non-technical audience the structure of a trained large language model (LLM), and the base and fine-tuning LLM training stages
\- Explain why compute, data and algorithms are key to training ML systems


_Spend 15-30 minutes writing your response_
Explain how models like GPT-4 are trained. Some questions to guide your thinking are below:
- What are the components of the AI triad, and what role do they play in training AI systems?
- Where do we get the data for large language models?
- What does it mean to train an AI system?
- What is the architecture behind modern AI systems?
- What happens during the training process?


Explaining how models get trained

_Spend 15 minutes writing your response._
In this exercise, you'll put an LLM to the test. Pick a chatbot of your choosing from the list below:
- [ChatGPT](https://chat.openai.com/)
- [Claude](https://claude.ai/)
- [Gemini](https://gemini.google.com/)
- [Microsoft Copilot](https://chat.bing.com/)
In your chatbot of choice, run the following prompts
- You are an advisor to a policymaker. Prepare a one-page policy brief for the policymaker on the AI Triad.
- Unfortunately, the policymaker is in a bad mood and is only accepting briefs in the form of poems. Please turn your brief into a poem, using the same formatting.
- The policymaker is too busy to read this! It’ll have to be a haiku.
Now, test it yourself! Prompt it to create new policy briefs, summarize ideas, write new poems, or do another task of your choosing. Once you're finished, reflect on how you think the chatbot performed. Did it do what you expected? Was it more or less capable than you anticipated? What mistakes did it make?


Using AI tools

Get a feel for how current AI models work! In this exercise, you'll try out several state-of-the-art models. We recommend playing around with 3-4 demos.
Before you start each demo, spend 1 minute trying to predict what you’re about to see or hear. How good do you expect it to be? What do you expect its shortcomings to be?
As you’re watching the demo, consider the following questions:
- Are the state-of-the-art AI models more advanced than you expected?
- How might these systems disrupt different industries?
- What capabilities might be coming soon?
- Based on today’s demos, what _can’t_ AI do?​
​
**`Part 1: Large language model`**
LLMs take text as input, and output more text. They're commonly used for tasks like chatbots. To complete this exercise, you'll need access to one. We recommend **[Bing Chat](https://www.bing.com/chat)** because it's free and doesn't require signup, but you can use any chat model you'd like such as **[OpenAI's ChatGPT](https://chatgpt.com/)**, **[Anthropic's Claude](https://claude.ai/new)**, **[Google Deepmind's Gemini](https://gemini.google.com/app)**, **[Inflection's Pi](https://pi.ai/talk)**, or **[Meta's Llama 3](https://ollama.com/blog/llama3)**.
Ask your chatbot for a recipe for your favourite food, but delivered as a rhyming poem.
1．What did it do well? What did it seem to struggle with?
Now ask it how gradient descent works, summarised in a few sentences. Ask it follow up questions if you want more information. (Tip: You can use AI models while on the course to help understand resources - but be aware that they sometimes make mistakes!)
2．Does this align with your understanding from the resources?
Now, ask your chatbot to calculate '13943642 × 1343' without using tools. Compare this result with a calculator.
3．Did it answer correctly? Why might it be difficult for an LLM without tools to answer these kinds of questions?
​
You can compare your results against **[our sample results](https://docs.google.com/document/d/11v2UxgUNOanub4DWiSIoOb5A_892zxX_K-HmDPuMXW8/edit#heading=h.qrtutjqgmx4v)**.
​
**Part 2: Text to image model**
Text to image models take text as input, and output an image. To complete this exercise, you'll need access to one. We recommend **[Bing Image Creator](https://www.bing.com/images/create)** because it's free, but you can use any image generation model you'd like such as **[OpenAI's Dalle-3](https://chatgpt.com/) **(integrated with ChatGPT), **[Midjourney](https://docs.midjourney.com/docs/quick-start)** or **[Stable Diffusion](https://clipdrop.co/stable-diffusion)**.
Ask your model for a picture of your favourite animal, playing your favourite sport on the moon.
4．What did it do well? What did it seem to struggle with?
​
**Part 3: Text to video**
Text to video models take text as an input and output an image. You can try **[Open AI's Sora](https://sora.com/)** (requires a free account, not available in some countries), **[Google DeepMind's Veo 2](https://deepmind.google/technologies/veo/veo-2/)** or watch **[this video](https://www.youtube.com/watch?v=1BDqfPEhsCA)** showing demos from several video generation models.
​
**Part 4: Text to Audio model**
Text-to-audio models takes text as an input and output as an audio file. We recommend **[Udio](https://www.udio.com/home)** (requires a free account to create songs, can browse examples without an account). You can also try the podcast generation features in **[NotebookLM](https://blog.google/technology/ai/notebooklm-audio-overviews/)**.
​
Ask your model to create a song using the following prompt format: _A [genre #1] [genre #2] song about [topic]. _Uncheck “instrumental”, and set the version to v3. It should take 30-60 seconds to create.
​
**Part 5: Robotics**
There are several humanoid robots powered by AI. Here are a few examples:
- **[Tesla](https://www.youtube.com/watch?v=DrNcXgoFv20)**
- **[Figure](https://www.youtube.com/watch?v=Sq1QZB5baNw)**
​
**Part 6: Defence systems**
AI has been used in wars **[for over a decade](https://disarmament.unoda.org/the-convention-on-certain-conventional-weapons/background-on-laws-in-the-ccw/)**. These demos show how far we've come with these systems:
- **[Palantir Gotham](https://www.youtube.com/watch?v=rxKghrZU5w8)** - an AI-powered defence system from 2021, used to help analyse data and make decisions
- **[Anduril Bolt](https://www.youtube.com/watch?v=EEXI6r08908)** - autonomous drones that can track and strike targets.
- **[DARPA's RACER ](https://www.youtube.com/watch?v=5t3-eXHC5Zs)**- autonomous tanks.


AI demo

AI Governance (2024-2025)

How AI Systems Work

The Promise and Perils of AI

Foundations of AI Governance

The AI Policy Toolkit

How AI is Governed Today

Governance at Frontier AI Companies

Proposals to Govern AI

Contributing to AI Governance