AI Took My Job: A glimpse of a post-workforce future.

As senior editor of a medical science journal, my job was to improve the readability of articles, liaise with authors, resolve content questions, and coordinate with production staff to publish the monthly issue. It was the sort of labor that Francis Fukuyama characterized as “white-collar work that involves symbolic manipulation.” But it also involved human skills, such as interpersonal communication, native linguistic ability, imaginative problem-solving and, as Samuel Johnson put it, “integrity of understanding and nicety of discernment.”

During my twenty-year tenure, the executive director and I elevated the production quality of the publication using the New England Journal of Medicine as a model. Our efforts took a while but proved successful. The journal grew in size, circulation, reputation, and relevance to clinicians. It gave me a sense of accomplishment to deliver a high-quality publication to an appreciative readership. And it felt good to make a living in a useful trade, specifically by bringing the latest musculoskeletal science to clinical practice.

Then, in 2021, the journal’s Board of Directors hired a new executive director, a tech enthusiast. In one of our first meetings with him, he requested that we get a bid from a Singapore-based AI-driven publishing platform, Straive. It didn’t set off any alarms at the time, because the journal’s Board of Directors reassured us when they hired him that they had no intention of outsourcing the journal’s production capacity, which had taken fifteen years to develop and had significantly driven its growth.

However, two weeks after the bid came in, and a day after we closed the March 2022 issue, each member of the production staff was called into a Zoom meeting and told that their services would no longer be needed. I’d been there for twenty years, the production manager for fifteen, the copy editor for ten, and the layout and graphics artist for five. We were each offered a measly $1,000 in severance, conditional on our signing a confidentiality agreement and staying another month to assist with training the AI. We all forfeited the $1,000, refused to sign the confidentiality agreement, and declined their request to help train the AI.

 

The Board’s rationale for laying us off was straightforward: machines are less expensive and time-consuming than human beings. At our parting, my colleagues and I were given the trite reassurance that the decision was strictly about finances and not our performance. Later, we learned that a critical factor was a cost-cutting bonus in the contract of the new executive director, which incentivized the swapping of higher quality for lower expenses.

My co-workers scrambled to get jobs with other academic publishers. The copyeditor was the only one who landed a better job as senior editor at a psychiatry journal. The graphic artist landed a lower-paying job at a scholastic textbook publisher where she did “a bit of everything,” as opposed to what she did best. The production manager got a job that she “did not love” and was eventually laid off again. She has since been doing temp and contract work.

I didn’t return to academic publishing. I’d been in the field for twenty-five years and wanted out. I opted to retire early and live a simple, frugal existence: writing daily for four to six hours, working in the garden, walking around my small city and talking to folks, doing fixit work on the house, spending time with my wife and kid, reading, etc. However bucolic that may sound, it was really a kind of rejection of the industry and the people who fed my colleagues and me to the wood chipper. It was a refusal to work for any organization that could degrade me and the legacy of my labor.

The journal now has a staff of two: an executive director and an advertising manager. An administrative assistant remained a while, until the empty office and heavy workload became oppressive. I imagine that AI-precipitated layoffs will leave many workplaces similarly desolate, the remaining workers burdened with the abundant tasks that only humans can competently do. When I was a temp in the early 1990s, during those recession-era corporate layoffs, I went into many a reorganized and downsized office. The atmosphere was often toxic with insecurity, resentment, and low morale. I can only imagine that AI-restructured workplaces will present similar scenes of a few angry, fearful, and overworked employees miserably hunkering down with temps in a half-vacant suite. Those who cannot immediately find permanent employment after being laid off will likely be fodder for the temp and gig economies. Bereft of financial stability, health insurance, paid sick leave, and legal protections—not to mention self-confidence and self-worth—many will fall into genteel poverty.

Two years after AI replaced me and my coworkers, the journal abandoned print publication. During my tenure, reader surveys had shown print to be the preferred format of over a third of the journal’s subscribers. The print edition was our premier product—the primary focus of our efforts and the model for the digital edition. Print afforded readers the luxury of holding a nicely designed and printed volume, reading at a leisurely pace, and pausing to reflect on and reread the text without distraction. I suspect that ditching printing and mailing wasn’t due to a lack of demand for print, which had always been fairly stable; it was done to generate another executive ka-ching!

I recently spoke to the production manager, who said that the website, the primary conduit of commerce and content delivery, is “in pretty bad shape.” I skimmed the articles and noticed a large number of usage errors and ambiguities—mistakes that would easily have been caught by a human copyeditor—as well as an abundance of trite phrases, corny puns, and solecisms that detracted from the content. The figures and graphics, which in my time would have been reconstructed by our graphic artist to fit the overall design and cross-checked against the raw data to ensure their integrity, were cut and pasted from the author’s original. Who knows how accurately they reflected the findings reported in the text. In short, the journal is a ghost of what it had been when human intention and care were there to ensure its quality, scientific integrity, and relevance to the readership.

According to the journal’s website, it has a quarter fewer subscribers than the human-staffed journal did. Its table of contents contains half as many articles. Its 2020-2024 IRS Form 990s show that its revenue went down from $1.7M to $1.5M but that its net income went up from $28K to $328K. That net income may look good at first glance, but it is solely due to cutting expenses deeply enough to outpace decreasing revenues. At the cost of lowering its production quality, subscription revenue, and circulation, the journal saved exactly what it would have cost to fund a human production team. Worse, its net profit, which for a nonprofit shouldn’t matter much, came at the expense of its primary mission: to disseminate research findings to as many clinicians as possible and thereby improve clinical practice and patient outcomes.

I am of course not the only human to be displaced by AI and to see the humanitarian mission of my organization swapped out for net profit. I expect that most of the companies who adopt AI in lieu of human labor will suffer reductions in quality, relevance, and reach. Companies with a skeleton staff, greater profit margins, and low overhead will eventually see their products and services lose market share. Innumerable accountants, statisticians, analysts, engineers, programmers, designers, project managers, and admins—the backbone of the country’s middle class—will be stripped of benefits, job security, health care, and financial security, and all for the sake of some nominal short-term profits.

It will take a while for these laid-off workers to recover from their anger and despair at that loss of agency. I spent a year fuming and ranting at the walls and wanting my revenge on something over which I had no control. My angry purgatory ended when I found something I’d written in a notebook just before I was laid off: a short list of activities of which an ideal day for me would consist. In my final unhappy days at the journal, I had designed a day that would meet my essential needs and enable me to live a healthy, meaningful life. A day that was repeatable and achievable within my present means. As I was re-reading this note, I realized that, since my layoff, I was already living my ideal day. From then on, I was more grateful than resentful for how I was living.

I still have dreams about working at the journal—frantically tapping at a keyboard without a screen and without understanding what is happening. When I wake up from these dreams, I’m always relieved to find myself where I actually am. To know that, despite the Faustian bargain AI represents and the ills many of us will suffer for it, a meaningful life is still attainable.

 

*Jan DiVincenzo is a freelance writer in Northern California.

 

Source: https://www.persuasion.community/p/i-lost-my-job-to-ai