Resources: Explosive growth (FT)

Exercises: Explosive growth (FT)

Resources: Automation (FT)

Exercises: Automation (FT)

Resources: Projected AI progress (FT)

Exercises: Projected AI progress (FT)

Resources: Distribution (FT)

Exercises: Distribution (FT)

Resources: Today's AI systems (FT)

Exercises: Today's AI systems (FT)

In this unit, we'll explore how AI could reshape labor markets and economic growth. We'll examine AI's impact through multiple economic frameworks, from task-based models of automation to aggregate production approaches. We'll analyze how different types of automation - cognitive labor, R&D, and physical labor - might affect wages, employment, and economic output.
Understanding AI's economic effects requires careful modeling of both direct displacement and indirect productivity effects. We'll explore key concepts like elasticity of substitution between tasks, bottlenecks to automation, and how growth effects can compound or be limited by structural economic factors. The unit emphasizes both theoretical frameworks and practical constraints that may affect AI's economic impact.
Keep in mind that predicting AI's economic effects involves significant uncertainty. Different models highlight different mechanisms and can yield varying conclusions. As you study these approaches, consider how their assumptions shape their predictions about AI's impact on labor markets and growth.

By the end of the unit, you should be able to:
\- Understand task-based models: jobs as bundles of tasks, and trying to figure out how automatable these tasks are on what timescales, and how substitutable they are
\- Model the impact of automating each of cognitive labour, R&D, and physical labour (via robotics):
\- Understand bottlenecks to automation in all three of the above


These exercises aim to get you to think about the models you have just read about in a holistic way. Many have been built with previous bouts of automation in mind, and are now being applied to predict the impacts of AI. We want you to think about where this might work well, and what it might be missing based on what you have read and discussed in the last two units. 

**Which of the models Trammel and Korinek present are the most realistic given the likely shape of AI technology? Which are the least realistic? (15 minutes)**

You may want to consider:
- What you have read about AI progress in the last two units
- The channels existing and future AI systems will most affect the economy
- Which of the channels is most sensitive to the strength of AI systems?



Model #1

These exercises aim to get you to think about the models you have just read about in a holistic way. Many have been built with previous bouts of automation in mind, and are now being applied to predict the impacts of AI. We want you to think about where this might work well, and what it might be missing based on what you have read and discussed in the last two units. 

**How much do you think automation from AI will differ from Acemoglu and Restrepo’s model built with previous bouts of automation in mind? (15 minutes)**

You may want to consider:
- What you have read about AI progress in the last two units
- What were some of the key characteristics of previous waves of automation were
- What are the principal assumptions Acemoglu and Restrepo make about how automation will evolve over time?



Economics of TAI Fast-Track

Exercises: Automation (FT)

Exercises