Resources: Projected AI progress

Exercises: Projected AI progress

Resources: Today's AI systems

Exercises: Today's AI systems

Resources: Distribution

Exercises: Distribution

Next steps: Programs

Resources: Automation

Exercises: Automation

Resources: Explosive growth

Exercises: Explosive growth

In this unit, we'll examine AI's potential distributional impacts across workers, regions, and countries. We'll explore how automation of increasingly complex tasks could affect wages and capital shares, with particular attention to how these effects might vary across different skill levels and economic contexts. We'll also analyze the methodological challenges in predicting these impacts and some concerning potential dynamics for developing economies.
The unit emphasizes both theoretical frameworks and empirical measurement challenges. We'll examine how different assumptions about task complexity distributions and capital accumulation lead to dramatically different predictions about wage trajectories. We'll also explore potential analogies to other economic phenomena, like the resource curse, that might help us understand AI's distributional implications.
As you study these approaches, consider how AI's impacts might differ from previous technological transitions, and what this means for both developed and developing economies. Special attention will be paid to the challenges of prediction and measurement in this rapidly evolving field.

By the end of the unit, you should be able to:
\- Understand the likely impacts of AI on the capital share
\- Understand possible impacts on wages and unemployment, within and across 
\- Understand the differential effects on high and low-productivity countries
\- Understand the open problems in the field


The distributional consequences of AI may be as important as growth ones for the “losers,” and could inherently limit growth themselves for political economy reasons or perhaps through economic mechanisms. We think this is one of the most promising areas for economists to publish cutting-edge research in, and hope this exercise can serve as a space for you to start considering your thoughts on this.

**What will happen to the distribution of income (and wealth) within countries, both across workers and factors? What might this mean for international income disparities? (30 minutes)**

You might wish to consider
- Household asset-holding heterogeneity
- Capital flows
- Migration
- Skill-bias of likely advances in AI systems
- Trade

We have very much focused on workers and automatable capital, yet wealthy people seem to have a strong preference for non-accumulable housing in central London and holidays on the French Riviera. This could limit the impact of AI.


Predicting Income Distribution

**What other non-accumulable factors (excluding human labour) might impact distribution and growth, and how? Give your best story for how each factor might become accumulable over time (10-15 minutes)**

You might wish to consider
- Income and wealth effects
- Directed technical change
- Regulation
- The economics of cities
- How the accumulability and relative importance of factors could change over time. 


Distribution and Growth

Finally, we want you to come to this discussion with some thoughts on directions in this field.

**List some open questions you have. (10 minutes)**

You might wish to consider
- How wide are your confidence intervals for the rate of growth?
- Which groups have we not considered distributional impacts on?
- How can we measure welfare gains and losses across a heterogeneous population?
- What further research is needed from the area, and how can we draw from existing economics?


Economics of TAI

Next steps: Programs