It would be fair to say that 2024 Nobel Prize winner Daron Acemoglu has been a bit of a sceptic about the impacts of generative AI (for example, see here). This scepticism is exemplified in a new paper forthcoming in the journal Economic Policy (ungated earlier version here). Acemoglu first notes that:
Some experts believe that truly transformative implications, including artificial general intelligence enabling AI to perform essentially all human tasks, could be around the corner... Other forecasters are more grounded, but still predict big effects on output. Goldman Sachs (2023) predicts a 7% increase in global Gross Domestic Product (GDP), equivalent to $7 trillion and a 1.5% per annum increase in US productivity growth over a 10-year period. Recent forecasts from the McKinsey Global Institute suggest that generative AI could offer a boost as large as $17.1 to $25.6 trillion to the global economy, on top of the earlier estimates of economic growth from increased work automation (Chui et al., 2023). They reckon that the overall impact of AI and other automation technologies could produce up to a 1.5–3.4 percentage point rise in average annual GDP growth in advanced economies over the coming decade...
Acemoglu then asks: "Are such large effects plausible?" His answer is a resounding "no". In making this assessment, Acemoglu relies on a relatively straightforward approach, which essentially boils down to working out the proportion of job tasks that will be affected by AI (relying on the earlier research I discussed here), how much they will be affected (in terms of potential cost savings), and how much of the economy those jobs make up. Overall, he estimates that:
...TFP [Total Factor Productivity] effects within the next 10 years should be no more than 0.66% in total – or approximately a 0.064% increase in TFP growth annually.
The effect of TFP growth on overall GDP growth rates depends on what happens to capital investment. Taking that into account, Acemoglu estimates that:
With these investment effects incorporated, GDP is also estimated to grow by 0.93–1.16% over the next 10 years. When I assume that the investment response will be similar to those for earlier automation technologies and use the full framework from Acemoglu and Restrepo (2022) to estimate the increase in the capital stock, the upper bound on GDP effects rises to around 1.4–1.56%.
Acemoglu then notes that AI will have bigger impacts on tasks that are easier to learn. He assumes that:
...productivity gains in hard tasks will be approximately one-quarter of the easy ones. This leads to an updated, more modest increase in TFP and GDP in the next 10 years that can be upper bounded by 0.53% and 0.90%, respectively.
Finally, Acemoglu argues that even these tiny GDP growth rate estimates likely overestimate the impact on wellbeing, because of potential offsetting negative impacts of AI, such as AI-powered social media. However, I think that point fails to take account of the fact that consumer surplus is not included in GDP growth (a real problem when you consider the contribution of 'free' (zero price) goods such as social media to GDP).
Nobel Prize winner Robert Solow once famously remarked that "You can see the computer age everywhere except in the productivity statistics". Clearly, Acemoglu is expecting something similar in terms of AI. This is a very public prediction of future AI-related growth, and it will be interesting to see if it is close to accurate over time.
[HT: Marginal Revolution, in April last year]
No comments:
Post a Comment