The MIT Technology Review article explores economist Daron Acemoglu’s views on the future of artificial intelligence and the broader economic forces shaping its impact. Rather than focusing on speculative predictions about superintelligence, Acemoglu identifies three practical developments that deserve close attention: workplace automation, the concentration of power among large technology firms, and the possibility of designing AI systems that genuinely assist workers instead of replacing them.
Acemoglu challenges the dominant narrative promoted by many Silicon Valley executives who describe AI as an inevitable engine of productivity and prosperity. He argues that the technology’s effects are not predetermined. The way AI is developed, deployed, and regulated will determine whether it benefits society broadly or mainly enriches a small group of corporations and investors. According to the article, he is particularly skeptical of exaggerated claims about near-term economic transformation and mass labor replacement.
One major concern involves what Acemoglu calls “so-so technologies,” systems that automate tasks without delivering meaningful productivity gains. He warns that companies may adopt AI primarily to cut labor costs rather than improve products or create new forms of value. In that scenario, workers could lose bargaining power while economic inequality widens.
The article also emphasizes the growing dominance of a handful of AI companies that control computing infrastructure, training data, and cloud platforms. Acemoglu argues that this concentration of power could limit competition and reduce democratic oversight of technologies that increasingly shape communication, employment, and public life.
Despite his concerns, Acemoglu does not reject AI innovation. He believes governments, researchers, and businesses still have an opportunity to guide AI toward worker-supporting applications that enhance human expertise instead of eliminating it. The article presents his perspective as a counterweight to both utopian and catastrophic visions of AI, emphasizing that economic outcomes will ultimately depend on human choices, institutions, and incentives rather than algorithms alone.