Artificial intelligence is expected to replace only about 5% of the work currently done by humans over the next decade, according to Nobel laureate Daron Acemoglu. This prediction was shared by Microsoft AI CEO Mustafa Suleyman, who argued that AI development should focus more on improving workers' capabilities than replacing them. Acemoglu's broader argument is that AI should be developed to complement human workers rather than replicate or replace them.
Suleyman shared Acemoglu's assessment in a post on X while introducing the first issue of The Humanist Review, a magazine published by Microsoft AI exploring the future of artificial intelligence. The post noted that 52% of Americans are worried about AI's impact on their jobs, but argued that the concern is "overblown" and is influencing how AI is being developed. This has significant implications for the future of work and the development of AI.
The economic impact of AI could be more gradual than some expectations, according to Suleyman's post. It suggested that AI isn't yet reflected in productivity statistics, and most firms using it aren't seeing real gains. The post estimated that AI could add roughly 1.5% to GDP over 10 years, not a revolution. This gradual impact is likely due to the need for companies to reorganize their operations around the technology.
The adoption of AI is being compared to the spread of electricity, with power stations established in New York and London by 1881, but only about half of factories and homes using electricity by the 1920s. AI adoption could be even slower as companies need to reorganize their operations around the technology. This slow adoption rate has significant implications for the future of work and the development of AI.
The post also cautioned against full automation even when AI systems become highly accurate. It noted that "even 99% accuracy isn't enough for full automation. The last 1% is the hard part." This suggests that AI systems need to be highly accurate and reliable before they can be used for full automation. The focus should be on "pro-worker AI," suggesting tools that make people better at their jobs.
Suleyman's post also flagged policy and market concerns, saying that "the US taxes labor at over 25% and capital at close to zero, which effectively subsidizes automation." It further noted that the seven largest technology companies account for 60% of the Nasdaq, arguing that "concentration crowds out new ideas." These concerns highlight the need for a more nuanced discussion about the development and deployment of AI.
The predictions and arguments made by Acemoglu and Suleyman have significant implications for the future of work and the development of AI. As AI continues to evolve and improve, it is essential to consider the potential impact on workers and the economy. By focusing on "pro-worker AI" and developing tools that complement human workers, it is possible to create a future where AI and humans work together to drive economic growth and productivity.
Key points
- AI may replace only 5% of human work over the next decade.
- AI development should focus on improving workers' capabilities rather than replacing them.
- The economic impact of AI could be more gradual than some expectations, adding roughly 1.5% to GDP over 10 years.