In the Me, Myself, and AI podcast from MIT Sloan Management Review, Nobel laureate Daron Acemoglu explains that, despite the hype, AI hasn’t delivered large measurable productivity improvements in the economy so far. While companies and governments have invested heavily in AI tools and infrastructure, most national and corporate productivity statistics do not yet show strong gains connected to AI adoption — a phenomenon reminiscent of past “technology productivity paradoxes.” Early studies suggest that AI might raise productivity modestly over the long term, but the immediate, dramatic effects many promised have not materialised in broad economic data.
Acemoglu stresses that much of the current AI deployment focuses on automating existing tasks rather than creating new tasks that significantly enhance human productivity. Technologies that simply replace human effort in routine work often result in limited economic value because they don’t expand what workers can do — they just shift who does it. Unless AI is used to create entirely new job categories or augment human work in meaningful ways, the net productivity benefits remain small.
Another challenge, he notes, is that today’s productivity measurements can be slow to capture the real effects of technological change. Historical examples, such as the introduction of personal computers and earlier digital technologies, show that productivity effects often appear with a substantial lag. However, Acemoglu warns that relying on future payoff alone isn’t enough; organisations need to focus on how AI is deployed so that it genuinely enhances labour productivity rather than simply reducing headcount or shifting costs.
Finally, Acemoglu points out that the distribution of AI benefits matters just as much as their magnitude. Even if AI increases aggregate productivity modestly, unequal access to those gains — such as favouring capital owners or high-skilled workers — could worsen inequality without broad societal improvement. For AI to live up to its promise of driving productivity growth that benefits many, it must be coupled with intentional organisational, economic, and policy choices that direct the technology toward tasks that augment human capabilities and create new economic opportunities.
