Three things in AI to watch, according to a Nobel-winning economist

A few months before being awarded the Nobel Prize in economics in 2024, Daron Acemoglu published a paper challenging the prevailing Silicon Valley narrative that AI would overhaul all white-collar work. He estimated that AI would provide only a small boost to US productivity and would not eliminate the need for human labor. He posits that while AI can automate specific tasks, many jobs will remain largely unaffected. This measured view stands in contrast to the pervasive discourse about an impending AI-driven jobs apocalypse.

Acemoglu acknowledges that the technology has advanced significantly since his initial predictions, yet data continues to support his stance that AI is not currently impacting employment rates or layoffs. He believes that the critical factor determining AI's impact on jobs is its ability to handle the complex orchestration between diverse tasks that humans perform effortlessly. He argues that AI agents, while advanced, are better suited as tools to augment specific aspects of a job rather than replacing entire human roles, as they still struggle with seamlessly switching between various functions.

One of Acemoglu's key concerns revolves around the increasing trend of AI companies hiring in-house economics teams. Companies like OpenAI, Anthropic, and Google DeepMind have recruited prominent economists, a move Acemoglu views with trepidation. He fears that these companies, driven by significant financial incentives, might influence research to promote their own viewpoints and generate hype, thus compromising the objectivity of studies on AI's societal and economic impact.

He also highlights the current lack of user-friendly AI applications that are as accessible and intuitive as past transformative software like Microsoft Word or PowerPoint. Acemoglu believes that for AI to have a truly seismic impact on the job market and the broader economy, it needs to develop more universally usable applications that workers can easily integrate into their daily tasks without extensive training or specialized knowledge.

The ongoing debate surrounding AI's impact is characterized by significant uncertainty. Despite the confident rhetoric from some quarters, the actual evidence regarding AI's effect on productivity and employment remains mixed and often contradictory. Acemoglu emphasizes this pervasive uncertainty as the most telling aspect of the current AI economy.