Keynote lecture by Daron Acemoğlu - 2024 Nobel Prize in Economics - "Progress at a turning point: Technology, power, and global challenges"

Author: Dolores Abuin, in collaboration with members of the cAIre-Google.org Project group, Ricardo Palomo (Coordinator), Jorge Cerqueiro, Frank Escandell, and José Javier Sesma.  

This article was produced by OdiseIA as part of the Google cAIre project. You can find out more about this initiative at the following link: Google Caire - OdiseIA

Taking advantage of Professor Daron Acemoglu's presentation during the Vigo Global Summit 2025, we had the unique opportunity to attend the event and make a productive comparison with our work already published on the cAIre Project website, within the Google.org Digital Futures Project.

As a professor at the MIT Sloan School of Economics, Acemoglu's analysis allows us to establish significant points of convergence with our conclusions from subproject 2.1 "Professions of the Future: New Opportunities and Requirements," while introducing a fundamental divergence in the diagnosis of labor transformation.

Points of convergence

The first convergence is the dual vision. Acemoglu articulates the crossroads between the path of Alan Turing (replacement) and that of Norbert Wiener (augmentation). We included this in our deliverable in the section on "Dual perception: risks and opportunities," which some experts identified as the core of the debate.

The second point of convergence lies in the skills of the future. Acemoglu's conclusion focuses on "learning to learn." This idea is reinforced in our study, both through surveys of the general public and those conducted with experts, who identified "critical thinking" and "adaptability and continuous learning" as the most crucial skills, above purely technical ones, with some emphasizing the "multitasking" nature of the professional of the immediate future. Both analyses conclude that the human value that translates into the market lies in skills, not in simple technical execution.

And the third convergence concerns regulation. Acemoglu's warning about the power of "Big Tech" and the need for "social and political" choice aligns perfectly with our section on "Governance, regulation, and applied ethics," which we identify as one of the main generators of new professions. Many of the future professional profiles we propose in our study will deal with rehumanization in the deployment of these technologies. "Performed by Humans" could become a whole market niche.

Points of divergence

This is where Acemoglu's thesis offers the greatest value to our work. 

Our report, based on the opinions of experts and the public, concludes that the impact of AI is an "inevitable transformation" and focuses on adaptation strategies (such as reskilling or horizontal learning and the creation of new roles).     

Acemoglu radically challenges this premise. He argues that the current technological trajectory (the Turing path) is not inevitable at all, but rather a choice.

The problem is not AI, but its direction. While our report analyzes the "impact" of AI, Acemoglu forces us to ask why AI has that impact. His diagnosis of the three obstacles (economic concentration, Hollywood ideology, and geopolitics) suggests that reskilling alone is insufficient if the direction of technological innovation is not changed.  

Acemoglu's solution is "pro-worker AI" (the "augmented electrician"). He offers us a new vision of the future of work that complements our report. The list of "new professions" in our deliverable (such as "Algorithm Auditor," "AI Solutions Architect," and "Data Curators") runs the risk of creating a new technological elite. Acemoglu reminds us that the real opportunity is not just to create new technical roles, but to use AI to increase the productivity, value, and wages of millions of existing jobs.

1. The arrival of AI

In his lecture, Nobel Prize winner Daron Acemoglu presents his thesis on the path of Artificial Intelligence as a social, economic, and ideological choice. It is a choice that, according to him, is leading us down the wrong path, with serious consequences for inequality, shared prosperity, and democracy itself.

Turing versus Wiener

Acemoglu argues that, since its inception, computer science has harbored two opposing philosophies about the relationship between humans and machines.

Alan Turing (automation). This is the dominant view today. It is based on the idea that the human mind is a "computing machine," a view also shared by Marvin Minsky and John McCarthy, the driving forces behind the 1956 Dartmouth Conference. The goal, therefore, is to create machines that mimic, compete with, and ultimately replace humans in their tasks. Success could be measured by parity. Can AI do a human's job equally well or better?

Norbert Wiener (augmentation). The other alternative view is based on the idea that machines should be tools that do things humans cannot do. This is why Wiener founded cybernetics, which is opposed to symbolic artificial intelligence. The goal is not to replace, but to augment and complement human capabilities. Success is measured by the expansion of our abilities (as did the mouse, hypertext, and the graphical user interface).

The technology industry, influenced by a wide range of factors, has almost entirely opted for Turing's path, generating what Acemoglu calls "uncontrolled automation."

2. Social consequences (inequality and democratic crisis)

The result of choosing Turing's path is clear to Acemoglu. It will lead to a huge increase in inequality.

This economic gap, he argues, is the direct fuel for the crisis of democracy we see today. When democracy fails to deliver on its promise of prosperity, populism, misinformation, and a loss of trust in institutions arise.

3. The solution is "pro-worker AI" 

Acemoglu proposes an alternative, "pro-worker AI."

His key example is not a radiologist, but an electrician. An electrician who has just started working cannot solve complex problems that an experienced one can. "Automation" AI (Turing) will try to replace both. "Pro-worker AI" (Wiener) would be a tool (perhaps in a portable device) that provides the beginner with real-time, data-driven information, allowing them to perform more sophisticated tasks and learn in the process.

This AI does not replace, but rather increases worker productivity. It is the "perfect antidote" to inequality, because a more productive worker capable of new tasks justifies a higher salary. This is the true promise of AI: not General Artificial Intelligence (GAI) that replaces us, but tools that double the productivity of doctors, nurses, teachers, plumbers, and laborers.

4. Why don't we have this AI?

If "pro-worker AI" is technically easier and economically superior for society, why don't we see it everywhere? Acemoglu points to three obstacles that prevent the direct emergence of such AI.

The economic obstacle (concentration). Unlike the schematization made by Amy Webb, professor at NYU Stern School of Business, on M.A.F.I.A.G. in the United States and B.A.T. in China, for Acemoglu the market is dominated by six or seven "Big Tech" companies. Meta and Microsoft's business model is not to "create tools for electricians in SMEs." Their models are based on Generative Artificial Intelligence (GAI), the sale of software to large corporations, and advertising, which encourages large-scale automation.

The ideological obstacle (Hollywood). We are culturally obsessed with the idea of AI competing with humans ("robots that kill us"), thanks to science fiction. This ideology attracts talent and funding to Turing's path, rather than Wiener's more "boring" but useful path. Ironically, this has been the same path by which the film industry is already clearly threatened.

The geopolitical obstacle (political weapon). AI has become a "political weapon" in a "zero-sum" race against China. The narrative "whoever dominates AI will dominate the world" justifies unlimited investment in automation, benefiting the "Big Tech" companies that promote it. The advent of Project Stargate by the Trump administration in early 2025, initially a billion-dollar initiative powered by OpenAI, Softbank, and Oracle, received a fleeting Chinese commercial attack with the disruption of the DeepSeek model shortly thereafter. The race for AI supremacy is not a new "Manhattan Project" but a new "Apollo Mission."

5. Conclusion

Acemoglu concludes with a serious warning about education. Acemoğlu criticizes tools such as ChatGPT for encouraging the pursuit of "expert answers" without the effort of the process, which threatens the most important skill of the future, "learning to learn."

The future of AI, therefore, should not be left in the hands of the few companies that control it. Daron Acemoglu's argument is that the "profession of the future" is not a technological destiny that we must predict, but a political, economic, and social choice that we must actively embrace. The great challenge is not technical, but civic.