AI Won't Eliminate Our Jobs -- It Will Expand ThemAI Won't Eliminate Our Jobs -- It Will Expand Them

Entry-level jobs and basic tasks are often derided as boring, basic or ripe for automation. But they're also valuable for helping us become better at what we do. And then, often, what we did gets automated out of relevance. Anyone who ever worked as an "HTML designer" in the mid-1990s can attest to this.

That AI is going to do what people used to do is not in doubt; this is what happens with every technological development -- think of spreadsheets (which helped one person do the tasks that used to take multiple bookkeepers and accounting clerks to do) and desktop publishing tools (which changed how publications went to press). But is AI is coming for our jobs? 

Last year, Washington University researchers Xiang Hui, and Oren Reshef, who teamed up with Columbia University's Luofeng Zhou published the study, "The Short-Term Effects of Generative Artificial Intelligence on Employment: Evidence from an Online Labor Market," in which they examined the impact gen AI had on creative freelancers who found work on Upwork. Hui, Reshef and Zhou looked at freelance writers and graphic designers on Upwork, citing specific attributes of ChatGPT as an LLM designed for text prediction and generation as a reason to look at text-heavy work. They found that freelance writers saw a 2% decrease in the number of monthly jobs and a 5% decrease in monthly earnings. In other words, not only did the number of gig opportunities drop, so did the pay being offered for the remaining jobs. The study also found that highly-ranked, high-earning freelancers saw the same dip in gigs and payments. In other words, being more skilled was no hedge against the job-market effects of gen AI.

Now, the gig economy is not a direct reflection of all professional classes and markets, and there have been other studies showing that instead of eliminating employees, gen AI actually makes workers better -- a Harvard Business School study found that consultants who were able to incorporate GPT-4  into their work were able to do 12% more work in 75% of the time, and their work was deemed to be 40% better than the control group work. The working paper's authors stressed that these effects held only to work "within the frontiers of AI capabilities," and if people used the gen AI tools outside those parameters, the AI-assisted work was substantially worse. 

So what are we looking at with these studies? It might help to look at how technology has affected jobs in the past. In 2015, a segment on NPR's "All Things Considered" examined how the electronic spreadsheet affected businesses -- and jobs. As journalist Jacob Goldstein said about Visicalc's impact: "Lots of bookkeepers and accounting clerks were replaced by spreadsheet software. But the number of jobs for accountants? Surprisingly, that actually increased. Here's why - people started asking accountants … to do more."

A year later, Boston University's Executive Director of the Technology & Policy Research Initiative, James Bessen, released a working paper on research examining how computer automation affects jobs and found, "Occupations that use computers grow faster, not slower. This is true even for highly routine and mid-wage occupations."

So if you're a manager or workplace strategist, you've got your talking points for how to introduce and manage gen AI tools in your workplace this year: This technology might take your job -- but it will definitely change it. How to help the workforce remain engaged and manage those changes will be an ongoing challenge.