Resources: Projected AI progress

Exercises: Projected AI progress

Resources: Today's AI systems

Exercises: Today's AI systems

Resources: Distribution

Exercises: Distribution

Resources: Automation

Exercises: Automation

Resources: Explosive growth

Exercises: Explosive growth

Welcome to the course! In this unit, we'll examine how modern artificial intelligence works and what it can do. We'll start with the fundamentals of deep learning - how neural networks learn patterns from data - and trace the field's rapid progression over the last decade. This technical foundation will help us understand why and how AI capabilities have expanded in recent years.
Today's AI systems show striking capabilities - from matching human performance on academic tests to generating realistic media to tackling complex research tasks. But their abilities are uneven, with some human-easy tasks remaining challenging while formerly difficult cognitive tasks become routine. Through concrete examples and benchmarks, we'll build an accurate picture of the current technological frontier.
Note that we've selected introductory materials for this first unit – more technical papers are available in the further resources for those interested in deeper analysis.

By the end of the unit, you should be able to:
\- Understand how modern AI systems learn using deep learning
\- Understand the history of deep learning and the progression of AI systems over the last decade
\- Explain what current AI systems are capable of across language, image and video generation, reasoning, robotics, and R&D


In this exercise, you’ll get a feel for what current AIs can do, by trying out models and watching 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. [Bing Chat](https://www.bing.com/chat) is 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). 

Some ideas for what to try:
- Some technical or conceptual question you recently worked with
- Some of the models allow uploading a PDF. How well can it summarise and critique a paper that you know well?
- Ask the chatbot for a recipe for your favourite food, but delivered as a rhyming poem.
- Ask it to summarise AI scaling laws. 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!)
- Ask it to calculate 13943642 × 1343 without using tools. Compare this result with a calculator. Did it answer correctly? Why might it be difficult for an LLM without tools to answer these kinds of questions?
​
You can see results for some of the above in [our sample results](https://docs.google.com/document/d/11v2UxgUNOanub4DWiSIoOb5A_892zxX_K-HmDPuMXW8/edit#heading=h.qrtutjqgmx4v) file.
​
**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.

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, or [this set of demos from Veo 2](https://www.youtube.com/watch?v=G9RDHs9nx04).
​
**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), or [Suno](https://suno.com/). 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.

(Or see [here](https://suno.com/playlist/34e6de43-774e-44fe-afc6-02f9defa7e22) for an ex-OpenAI researcher’s musical version of a Kipling poem)
​
**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

Economics of TAI

Understand how today's AI systems work and what they can do

Today's AI systems

Map the trajectory from current systems to transformative capabilities

Projected AI progress

Analyze economic frameworks for AI-driven labor market transformation

Automation

Explore AI's potential to accelerate economic growth

Explosive growth

Track the distributional impacts of widespread transformative AI adoption