The Economist recently published an article on Nobel laureate Daron Acemoglu [who won the 2024 prize along with Simon Johnson and James A. Robinson]. The article engages with the economist’s academic work disingenuously, quoting snarky comments made by unnamed economists to support an argument that his influential research makes sweeping, unfounded claims.
Massachusetts Institute of Technology (MIT) Professor Acemoglu’s research linking institutions to long-run prosperity—the work that the Nobel committee honoured two years ago—is portrayed as little better than glib, with his pessimism on artificial intelligence (AI) and preference for steering technological change through public policy resting on dubious assumptions and an underestimation of AI’s potential economic benefits.
Acemoglu has been a tireless advocate for AI policies that are job-centred, humanistic and compatible with liberal democracy. His proposals are probably anathema to titans of the AI industry like Elon Musk and Mark Zuckerberg. For them and other prominent figures, workers and democracy may seem like irritating obstacles on a path to ever-greater wealth and power.
That the academic credentials of a researcher who has been publishing in peer-reviewed journals for more than 30 years are being questioned [could be interpreted as] frenzied defence of the faith by offended true believers.
Obviously, there is nothing improper about criticizing Acemoglu’s scholarship and reckoning with his ideas, no matter the timing. Nobel laureates deserve scrutiny just like everyone else. But why does an article ostensibly reassessing the quality of Acemoglu’s academic work end by attacking his influence over AI policy?
That structure is revealing. One “well-known economist” accuses Acemoglu of using models to support failed “populist policies”; another is quoted on his genius. It then tours old controversies over settler mortality, institutions and technology before arriving at AI.
Only there does the article disclose the stakes. These academic disputes, it says, “could end up mattering a great deal” because Acemoglu worries about AI’s effects on jobs and democracy.
That pivot deserves more attention than the ‘econometric archaeology’ preceding it. The Economist has a recognizable position on AI policy. In a 2023 Leader, it rejected both laissez-faire and maximal precaution, advocating a “measured approach” on which further rules could later be built.
In July 2026, another Leader called for better regulation while warning that America should not “imprison frontier AI” or lock others out of the technology.
That is, no doubt, a defensible position: regulate risks, but preserve innovation, diffusion and strategic advantage. It is also broadly congenial to the interests of the AI giants whose valuations depend on rapid deployment and relatively permissive rules.
Acemoglu’s agenda is different, and much more disruptive of the plans of the architects of artificial intelligence.
With his MIT colleagues David Autor and Simon Johnson, he has argued that the direction of AI is a policy choice, not a technological inevitability. Their 2023 “pro-worker AI” programme proposed equalizing the tax treatment of labour and equipment or algorithms, limiting intrusive workplace surveillance, strengthening workers’ voice, funding human-complementary technologies, building public AI expertise also and steering adoption in education and healthcare.
The follow-up 2026 programme of the MIT trio goes further, adding antitrust enforcement and intellectual-property protections for workers’ expertise. The objective that Acemoglu and his colleagues have set is not just safe AI, but artificial intelligence designed to increase the value of human capabilities rather than automate them away.
On 21 August, Acemoglu published a very clear statement of this policy path in the ‘World View’ section of the well-known science journal Nature.
This agenda directly questions who sets the technological trajectory and who captures its rents. Against that background, diminishing Acemoglu’s academic standing could have an obvious political effect: it could weaken the authority of one of the world’s most prominent economists arguing for a worker-centred approach to AI’s development.
The more interesting charge is ideological. If disagreement with Acemoglu’s policy prescriptions can turn into snide questions about the merits of his scholarship and academic “stardom,” it risks the sabotage of a debate that should be conducted openly and honestly.
Whether every empirical finding by Acemoglu from 2001 survives re-estimation is not as important as whether AI policy should accommodate the firms building the technology or give workers and the wider public greater power over what gets built. That is the argument that cheerleaders of AI should be having with Acemoglu. The debate over AI is hugely consequential. ©2026/Project Syndicate
The author is professor of economics, Tel Aviv University, and member, Centre for Macroeconomics, London School of Economics.