How AI is reshaping the workforce

Artificial intelligence (AI) is undoubtedly having an impact on the workplace. But what kind of impact precisely?

A study last year from IESE found that managerial jobs were actually growing, as organizations felt they needed more managers to oversee both the rollout and deployment of AI in the organization.

This was interesting, as a French study into the impact of robotics on the workforce showed that it actually required "fewer" managers. The picture has been further muddied by research from UC Berkeley, which set out to understand the wider impact AI was having on the workplace.

Changes in management

The authors highlight that the earlier adoption of AI between 2010 and 2018 coincided with a flattening out of management structures. They suggest that successful AI deployment needs large amounts of data and computing power, but also the right human capital. The first two are difficult to change, which prompts many companies to focus on the latter ingredient.

They looked at data from a few hundred million resumes from 2010 to 2018 and created an org chart for thousands of companies. They then compared this with data from around 180 million job listings.

Next, they attempted to understand the likelihood that any given employee would be working on AI. They paired up machine learning, natural language processing, artificial intelligence, and computer vision, with thousands of skills used on both resumes and job postings.

“Once we had that, we could simply count jobs at each company and see how that affected the rest of the workforce,” the researchers explain. “If a firm of 1,000 people hired five new AI employees, what happened to the other 995 people?”

Workplace changes

Across the board, they found that employment went up, with companies that invested in AI tending to innovate more and therefore bringing in more employees to do so. This chimes with previous research looking at the impact of investment in robotics. More investment made firms more efficient (and profitable), which coincided with more jobs (I'll get to whether that applies in the gen AI age later).

The composition of the workforce changed, however. There was a premium on those with at least a college education, with those jobs increasing at broadly the same rate as the jobs for those without a college degree fell. There was also a shift towards jobs requiring a STEM education, with those jobs replacing those with social science degrees.

There was also a hollowing out of management, with companies that invested in AI often investing in smart people who didn't need so much oversight by middle management.

Obviously, correlation doesn't mean causation, so the researchers were careful to link these workforce changes with investments in AI. They did this by exploring firms that had a strong link with a university, and especially universities that had a strong track record in AI, such as the University of Toronto.

“We were able to leverage which universities had these strong AI hiring networks before the commercial wave hit around 2010,” they explain. “We found that, when this shock hit, the firms that had pre-existing connections to those AI-strong universities but were otherwise like their peers were able to hire AI workers and establish AI teams more easily. From those early AI teams flowed the other effects on workforce composition.”

A positive picture

Now, it's justifiable to say that this is largely something that we've already known. After all, the dire predictions of widespread layoffs from automation had not materialized at all in the decade since Carl Frey and Michael Osborne made them in 2013.

What the situation is like since the release of ChatGPT in 2023 is arguably more important than exploring the situation in the decade prior to that. For instance, early data suggests that AI leads to industry concentration, with a small number of superstar companies dwarfing rivals, not least because they have the highest quantity of data, computing power, and talent needed for AI to succeed.

That doesn't necessarily translate into more jobs, however. For instance, this year has already seen Alphabet make significant layoffs in its cloud and Android divisions, despite cloud being a key driver of the company's profits, with revenue growth of 30% over the previous year.

If one of the undoubted leaders in AI is making significant efforts to shed workers, the notion that AI adoption means a bright future for workers doesn't seem guaranteed at all.

Ultimately, while early AI adoption was linked to job growth, the post-ChatGPT era presents a different landscape—one where industry concentration and strategic layoffs raise new concerns.

Sure, tech firms might not be going all Klarna and shouting from the rooftops that they're automating a large number of jobs, but with Elon Musk bullishly doing just that across the US government, it hardly seems like Silicon Valley can breathe any easier than the rest of us.