Two years later, Acemoglu’s measured approach has not resonated. Talk of an AI jobs apocalypse is popping up everywhere, from Senator Bernie Sanders’ rallies to conversations overheard in line at the grocery store. Some previously skeptical economists have gotten more It opens To the idea that something seismic might come with artificial intelligence. Candidate for Governor of California He said Last week, he announced that he wants to tax companies’ use of AI and pay compensation to victims of “AI layoffs.”
On the one hand, the data is still in favor of Acemoglu; Studies have repeatedly found that AI does not impact hiring or layoff rates. But technology has advanced little since his cautious predictions. I spoke with him to find out if any of the recent developments in AI have changed his thesis, and to find out what has him concerned these days, if not the impending general artificial intelligence.
Artificial intelligence agents
One of the biggest technical leaps in the field of artificial intelligence since Acemoglu’s research has been agentic artificial intelligence, or tools that can bypass chatbots and operate on their own to complete the goal you set for them. Because they can work autonomously rather than just answer questions, companies are increasingly promoting agents as a one-to-many alternative to human workers.
“I think this is just a losing proposition,” Acemoglu says. He believes that agents are better viewed as tools to enhance specific parts of someone’s job rather than as tools flexible enough to handle a person’s entire job.
One reason has to do with all the different tasks that go into the job, something Acemoglu has been researching in his work on AI. since 2018. For example, an x-ray technician handles 30 different tasks, from taking a patient’s history to organizing archives of mammogram images. A worker can naturally switch between formats, databases and business methods to do this, Acemoglu says, but how many individual tools or protocols will AI require to do the same thing?
Whether or not agents will enhance the impact of AI on jobs will depend on whether they can eventually handle the coordination of tasks that humans do naturally. AI companies are in a fierce competition to prove that their AI agents can operate autonomously for longer periods than ever before without making mistakes, sometimes exaggeration The results – but Acemoglu says many jobs will be saved from AI takeover if agents cannot seamlessly switch between tasks.
New hiring spree
For many years, major technology companies have been offering amazing salaries to hire AI researchers. But I asked Acemoglu about a different hiring spree I’ve noticed: AI companies are all building internal economics teams.
OpenAI has hired Ronnie Chatterjee from Duke University in 2024 as its chief economist Announce Last year, Chatterjee will work with Jason Furman — a Harvard economist and former advisor to Barack Obama — to research artificial intelligence and jobs. Anthropic has It was held A group of 10 leading economists to do similar work. And just last week, Google DeepMind announced that it had hired Alex Imas, an economist from the University of Chicago, to be its “director of economics for artificial general intelligence.”
