I implore you: stop saying AI is not taking jobs

When I ordered my Waymo on a chilly San Francisco morning, I didn’t know what to expect. I was in SF for work and I was nervous to try the autonomous vehicle, but my curiosity won the day. The car arrived - a Jaguar, no less. I fumbled with the handle, opened the door and settled into the back seat. After an introduction from a personless voice, I hit “start ride” and suddenly I was being being chauffeured to the GitHub office. NOBODY was DRIVING. I repeat: Zero other humans were in the car with me. It was, in a word, luxurious. I cut entire paragraph from this article that was essentially an ode to the Waymo. I loved everything about the experience and have shared this broadly since, to whomever might listen. But I’m an AI ethicist, and my love of innovation is checked and balanced by my consideration of the impact innovation has on very real people. Every time I booked a Waymo, I thought through the downstream effect the driverless cars have on the people of San Francisco who drive for a living - taxi, Uber and Lyft drivers, among others.

I broke with my beloved Waymo when, in the midst of my San Francisco trip, I jetted to Dallas for 18 hours to give a talk at a conference. In the talk just before mine the speaker said something that caught my attention. “We all know that AI isn’t taking jobs.” The first time they said it, I was certain it was an accident. Surely, I thought, they don’t mean to paint with such a broad brush. But they kept saying it - six times in a 25 minute presentation. I’ll admit, I was, and still am, surprised that this person built an entire presentation on a statement that was not only broad, but false, and dare I say, dangerous. I, myself, had just come from San Francisco where driverless cars carted me hither and yon across the city. Can we honestly say that AI was not taking jobs when AI-infused autonomous vehicles are directly competing in a market that not 5 years ago required one human driver for every car ride?

After finishing my own presentation, the statement from my predecessor was still nagging me, so I did some research. Sure, there are articles that claim that “AI is not yet killing jobs”, but these articles make a fatal mistake: they assume that (because of generative AI) white-collar jobs are the only jobs that could be impacted, or worse, they assume they’re the only jobs that matter. With the pace of AI innovation, these are narratives we can ill afford. Please hear me: I am not claiming that the robot apocalypse is here and all jobs are on the chopping block. Not every job is at risk, and how AI will impact a job depends on the utility of AI in that industry. There are even papers detailing the way implementation of robotics in labor jobs in China has actually “exerted a promotional effect” and “relatively improves the job share of women and workers in labour-intensive industries,” which is great news.

But positive signals in some geographies and industries does not mean that AI is not taking jobs. AI is certainly replacing humans in a variety of industries. Waymo is one example of that. Other examples include Wendy’s restaurants. The US based chain has implemented AI voice-ordering technology in the drive-thru in 36 locations, and is actively expanding the implementation. Levi’s began using AI fashion models to model for their website rather than humans. Coca-Cola just made their culturally beloved holiday commercials entirely with AI this year. Klarna very openly announced they would lay off much of their support staff because AI could perform customer support tasks better, faster, and cheaper. Farmers are adopting food processing robots, among other AI, which can do more, faster. Video game actors are still on strike after video game makers refused to explicitly state that the actor’s work would not be used to train AI which would eventually be able to replace the actors. Copywriters have reported job losses due to the launch of ChatGPT.

Each of these cases has a set of pros that accompany the cons, and I’m not claiming that there aren’t good - even great - uses of AI. But as humans, I think we often draw conclusions based on our own reality. So if we aren’t losing our job to AI, then it must not be the case for someone else. I like to remind people that just because something isn’t your reality doesn’t mean it isn’t the reality for someone. It’s true that relative to the number of jobs on the planet, AI’s displacement of workers has been minimal. But to the people who have been impacted, and to the people who will be directly or indirectly impacted by AI’s broad utility and availability, the displacement is painful and complicated. If we continue to peddle a dangerous narrative that “AI isn’t taking jobs” then we defer the very important conversation around what needs to be done to ensure those whose jobs will be eliminated due to AI have the necessary skills and resources to adapt in the job market. Who bears the responsibility for that upskilling? Is it AI companies? Companies adopting AI? The individual impacted? I’ve heard many a suit say, “It’s the person’s job to make sure they’re up to speed on AI.” But what if those broadly impacted by AI’s employment are some of the 2.6 billion people globally who are digitally excluded - lacking consistent access to technology, WIFI, or the digital literacy to meaningfully use those tools?

There’s a lot to be considered, so what do we do? Movement toward a compassionate view of AI’s presence in the job market is, in my opinion, threefold: first, we must stop saying that “AI isn’t taking jobs.” Second, we must adopt an understanding that AI’s impact on the future of jobs is essentially industry agnostic and hitherto unknown. Third, we must choose to meet each AI-powered layoff with empathy, and encourage companies and governments to consider upskilling resources for all those negatively impacted.