Artificial Intelligence on the Job - What’s Happening?

Although AI has not clearly been eliminating jobs, a lot has been going on with it at work.

Its effect on the up-and-down popularity of working from home hasn’t been moving in the direction many would expect, as “Return to office gaining momentum as AI reshapes corporate strategy” (Arabelle Bennett, Fox Business, November 2nd).  Per Newmark, “a global commercial real estate advisory firm that counsels Fortune 500 occupiers and major landlords on leasing, strategy and transactions,” “artificial intelligence is driving a surprising surge in office demand.”  That company’s president says these companies are “making bold moves” to “re-skill” employees “and have them retrained in artificial intelligence.”

Speaking of up and down, we next got “More!  More!  More!  Tech Workers Max Out Their A.I. Use” (Kevin Roose, The New York Times, March 20th).  As “at tech companies like Meta and Shopify, managers have started to factor A.I. use into performance reviews, rewarding workers who make heavy use of A.I. tools and chastening those who don’t.”  “An engineer at Open AI processed 210 billion “tokens” - enough text to fill Wikipedia 33 times - through the company’s artificial intelligence models over the last week,” and “at Anthropic, a single user of the company’s A.I. coding system, Claude Code, racked up a bill of more than $150,000 in a month.”  “Some tech companies, including Meta and OpenAI” had “internal leaderboards that show how many tokens… each worker consumes,” and on which “employees compete(d).”  The logical worker response is for them to use AI nonconstructively, which would put an end to this rather shortsighted practice.  Indeed, on June 18th, less than three months later, the Times printed Eli Tan’s “Tech Workers Maxed Out Their A.I. Use.  Now They’re Trying to Minimize It.”; the change came from “the bills from companies, like Anthropic and OpenAI, that provide A.I. tools - and they were not cheap.”  Going from almost mandating it to putting “some monthly limits on A.I. coding tools” - was that embarrassing, or just the cost of learning?

In what might be another natural outcome of such maxing out, “Meta’s Embrace of A.I. Is Making Its Employees Miserable” (Kalley Huang, Eli Tan and Kate Conger, The New York Times again, May 8th).  In order for that organization’s management officially “to capture employee data so Meta’s artificial intelligence models could learn “how people actually complete everyday tasks using computers,”” it gathered “what employees typed into their computer, how they moved their mouse, where they clicked and what they saw on their screen”; as a result, “many workers immediately revolted” and “blasted the tracking as a privacy violation, calling it antisocial and callous.”  In conjunction with substantial announced upcoming layoffs, many workers there were showing “anger and anxiety.”

Specifically, “What Are A.I. Agents Actually Doing?” (Cade Metz, The New York Times, June 4th).  “A San Francisco start-up called Arena, which tracks hundreds of thousands of artificial intelligence users, is trying to take some of the mystery out of what” tasks it is performing.  It found that 17% in “agent mode” were “code-writing,” 10% were for “research,” and almost as many were for making images, creating “documents like graphs and spreadsheets,” and to “brainstorm ideas.”  Significant shares also were spent on “creative writing or tutoring and education,” along with “code debugging” and “chatting.”  A wide variety.

Yes, it seems clear that “We’re Only Starting to Grasp the Pitfalls of Using A.I. at Work” (Noam Scheiber, The New York Times, June 29th).  Although “at a conference where two human resources executives said that treating A.I. agents like real employees was a way to increase productivity and to put their companies on the cutting edge,” “in an experiment involving dozens of companies with A.I. employees, the researchers found that managers tended to vet documents less carefully when told an A.I. employee had produced them,” and they “missed errors that other managers caught when told they were vetting the work of a human.”  As AI models “tend to favor work produced by artificial intelligence” causing a “potentially consequential form of implicit ‘anti-human’ bias,” and AI models do not often “cooperate and seek win-win outcomes” as people do, we can question whether businesses are using too much AI.  

Should companies slow down?  That would be hard for many to do, but it might be the best choice.  As these stories show, the race, even when it is about implementing artificial intelligence, is not always to the swift.