Field report SW-90 / N1 · sheet /operations

Operations & RestructuringFiled from the territory

Worker Backlash Deepens as 67% Predict a Worse Employee Experience

Worker sentiment turned sharply against AI: 67% expect a worse workplace, usage trails executive enthusiasm and costly errors are driving sabotage and resistance.

Grid
SW-90 / N1
Surveyed
November 19, 2026
Reading
5 min read
Correspondent
Priya Raman
‘No one asked for this’: AI is being pushed on people—and they’re sick of it
Fig. SW-90 / N1‘No one asked for this’: AI is being pushed on people—and they’re sick of it — AI-generated

Field notes

  1. 67% of June Ipsos respondents said AI would worsen the worker experience.
  2. Glassdoor’s positive AI-review share fell from 81% in 2019 to 43% this year.
  3. 69% of workers said their employers encouraged AI agents, but only 16% used them.
  4. 32% of US hiring managers who cut positions for AI reinstated the same or similar roles.
  5. Gartner predicts half of companies making AI-driven staff cuts will rehire by 2027.

By Priya Raman

Employers face an AI backlash: 67% of respondents to a June Ipsos poll said the technology would worsen the worker experience. Glassdoor’s positive AI-review share fell from 81% in 2019 to 43% this year. Pew Research found 52% of Americans were more concerned than excited, up from 37% in 2021.

Companies market AI as a solution, but employers urge workers to embrace it while cutting workforces through automation. Mass layoffs, environmental effects, privacy concerns and cultural “sloppification” compound the resistance.

On X, one user said nobody had asked for Google to get worse, computer hardware costs to rise or human customer service to be replaced: “AI made it happen.” Science-fiction, fantasy and mystery author Sarah Hoyt, who does not hate AI, is “begging all the crazy people out there to stop pushing AI where no one asked for it.”

Jiemin Tina Wei, an assistant professor of history at Florida International University, sees the closest analogy in the Industrial Revolution’s replacement of physical labor. AI instead replaces “brain work,” rather than manual labor, she said, and the industries allegedly wiped out are the “‘low-hanging fruit’ of white-collar work.”

Why are productivity gains failing workers?

Louis Hyman, a Johns Hopkins historian of labor and the American economy, calls the mismatch a “productivity paradox”: a contradiction also seen in the computer boom of the 1980s, when heavy investment did not necessarily improve workers’ jobs.

Frequent updates compound it. Hyman said: “Every time you learn how to do something, suddenly they break it, and you have to relearn it.” With AI, it feels “as if all software is being updated, all at once, everywhere.” He also said some workers’ use of AI tools remains unclear.

Hyman uses AI to write software and automate “boring tasks,” including a script rotating thousands of archive photos right-side-up. Yet “most people are not wanting to sit down and write Python code to automate their job. They just want to do their job.”

Replacement remains the fear. “I’m a tenured professor—I literally can’t be fired,” Hyman said. Automating a tedious task gives him more time for interesting work, “but that’s not true for all people, right?”

What happens when guardrails fail?

A lack of adequate guardrails has produced costly errors. One AI tool deleted a software company’s entire database; another was blamed for deleting an edited season of a Swedish reality show weeks before its premiere.

AI scribes inserted inaccurate information into patient records, raising liability questions over resulting medical errors. According to the MIT AI Risk Initiative, misreads by law-enforcement identification tools caused wrongful arrests and injuries. Other AI-generated errors appeared in court documents, federal reports, news stories and fast-food drive-thrus.

Radio host Adam Xii, after Anthropic hallucinated details in its defence during a music copyright lawsuit, said: “The main point I’ve pushed and will continue to push re: AI is ‘no one asked for this.’” He said ubiquitous tools require a “why,” because “AI tools are meant to generate more wealth for the wealthy at the expense of creatives.”

Convenience now comes with checking work that can create new human errors. The proliferation of “workslop,” or low-quality machine-generated content, prompted one technology company’s CEO to impose a full workplace ban, Inc. reported this summer.

Wei called the burden “annoyance at the proselytizers at work—or the people who are heavy adopters—creating a lot of extra work for the people around them.” She said content moderation and data annotation can be “extremely mentally taxing.”

She cited research finding that “good” challenging work is difficult and interesting to a degree, but checking output means searching for every small error and can be boring. “You’re not producing knowledge, but merely supervising the production of words, knowledge, content,” she said.

Are workers sabotaging AI plans?

Resistance appears in workplace behavior. Nearly one-third of workers surveyed had tried to sabotage their companies’ AI strategies; the share rose to 44% among Gen Z.

Section, an AI planning and upskilling company, found that 69% of workers said their organizations encouraged AI agents, meaning autonomous task-performing systems, yet only 16% used them. Thirty-eight percent used other AI features.

Microsoft nevertheless promoted Edge’s Copilot Mode for Business with: “We heard you wanted Copilot Mode at work.” A technician replied: “No, you heard wrong. Literally no one asked for all this AI. In fact, everyone wants to know how to remove it.”

Can employers turn AI into a retention tool?

Corporate arithmetic shows the cost of treating automation as a substitute for people. Robert Half found that 32% of US hiring managers who eliminated positions for AI reinstated the same or similar roles. Gartner predicts that, by 2027, half of companies reducing staff because of AI will rehire people for similar work.

Wei tied that reversal to labor history. She cited organizer Bill Haywood’s maxim that “the manager’s brains are under the workman’s cap,” arguing that workers hold the knowledge to run daily operations while managers seek to reclaim it. Some employees, she said, experience mandated AI as training for “their final, automated replacement.”

Hyman’s prescription is to empower employees, not force an “adapt or die” response. Organizations should “create organizational conditions that let you replace yourself, and then focus on the work that needs to be done,” he said, because “the amount of work to be done is not finite, right?”

He hears nearly weekly that historians will be replaced. “So, yeah: it depends on how valuable you think getting reliable information and history is.” If OpenAI or Anthropic supplied only history or journalism, the future would be an “Orwellian nightmare,” “not a great road to go down.”

For employers, the evidence favors a different operating model: empower workers rather than train their replacements, and account for verification work, renewed hiring and organized resistance.

via fastcompany.com/rss (Source)

Share this article:

More from Priya Raman

Priya Raman

Show full bio

Senior reporter covering technology and product news in corporate strategy and chief-executive leadership at CEO Frontier.

450 articles

Filed under

Adjacent sheets

« Previous articleNext article »

End of sheetSW-90 / N1 · November 19, 2026