Two years later, Acemoglu’s measured grip has not held. There is talk about the apocalypse of AI jobs everywhere from Senator Bernie Sanders’ rallies to conversations overheard in line at the grocery store. Some economists who were previously skeptical have found more open The idea that something seismic could happen with AI. California gubernatorial candidates Said Last week he wanted to tax corporate AI use and pay victims of “AI-driven layoffs.”
On the one hand, the data is still in Acemoglu’s favor; Studies repeatedly show that AI is not affecting employment rates or layoffs. But technology has come a long way since his cautious predictions. I spoke to him to understand whether any of the latest developments in AI have changed his thesis, and to find out what, if not imminent AGI, worries him these days.
AI Agent
One of the biggest technological leaps in AI since Acemoglu’s paper is agentic AI, or tools that can go beyond chatbots and act on their own to accomplish a goal you give them. Because they can work independently rather than simply answering questions, companies are increasingly introducing agents as one-to-many replacements for human workers.
“I think it’s just a losing proposition,” Acemoglu says. He believes that agents are better thought of as tools to enhance particular parts of one’s work rather than a tool flexible enough to handle a person’s entire work.
One reason for this has to do with all the different tasks involved in the job, Acemoglu has been doing some research on AI in his work. since 2018. For example, an X-ray technician performs 30 different tasks, from recording patient histories to organizing the collection of mammogram images. Acemoglu says a worker can naturally switch between formats, databases, and work styles to do this, but how many individual tools or protocols would AI need to do this?
Whether agents will supercharge AI’s impact on jobs will depend on whether they can eventually handle the orchestration between tasks that humans naturally perform. AI companies are in a fierce competition to prove that their AI agents can work independently for long periods of time without making mistakes, sometimes exaggerating the results – but Acemoglu says many jobs will survive AI takeover if agents can’t easily switch between tasks.
Competition for new appointments
For years Big Tech has been offering great salaries to recruit AI researchers. But I asked Acemoglu about a different hiring spree I’ve seen: AI companies are all building in-house economics teams.
OpenAI appoints Ronny Chatterjee from Duke University as its chief economist in 2024 announced Last year, Chatterjee will work with Jason Furman, a Harvard economist and former adviser to Barack Obama, to research AI and jobs. Anthropic has called A group of 10 leading economists to do similar work. And just last week, Google DeepMind announced that it had appointed Alex Imas, an economist at the University of Chicago, as “Director of AGI Economics.”
