Glassdoor just published data that should worry any manager who assumed employees would come around to AI once they got used to it. The opposite happened. Positive comments about AI in company reviews sat at 81 percent back in 2019. By the middle of 2026, that number had fallen to 43 percent, and negative comments now outnumber positive ones. The shift happened fast: as recently as 2025, sentiment was still 55 percent positive. One year changed the direction entirely.
The complaints are not mainly about robots taking jobs, though that fear shows up too. The bigger pattern is about how AI got introduced. Workers describe being handed tools with no training, being told to hit productivity targets nobody explained, or being monitored by automated systems that feel like surveillance. Insurance claims staff, writers, and customer service workers are the most negative groups, often blaming management for forcing buggy software on them. Meanwhile executives and software architects, the people who decide which tools get bought, stay largely positive.
That gap between who picks the tools and who has to use them is the real story here. It also explains why bigger companies score worse: at firms with more than 10,000 employees, 67 percent of AI comments are negative, compared to 51 percent at small companies. Bigger organizations tend to roll out AI top down, with less room for workers to push back or shape how it gets used.
The generational split adds another layer. Gen Z women are the most skeptical group in the entire workforce, with only 21 percent positive comments, compared to 42 percent for Gen Z men. Gen X, the generation with the most seniority and job security right now, is the most upbeat at 47 percent. Confidence about AI tracks closely with how safe someone feels in their job, not with how much they actually use the technology.
This sentiment shift is not happening in a vacuum. Tracking firms have found that more than half of layoff events announced in 2026 cited AI or automation as a factor, affecting well over 170,000 workers. Separately, a Gartner study of 350 companies found that the firms cutting the most jobs saw no better financial results than the ones that did not. Even OpenAI's own chief executive has said publicly that some companies are blaming AI for cuts they would have made anyway. Workers reading headlines like that while getting a new chatbot dropped into their workflow are not being irrational when they get defensive.
There is a useful split hiding inside the negative numbers. Ten percent of negative comments actually complain that their company is too slow to adopt AI or gives them the wrong tools. Those workers are not against AI, they are against bad execution. That is the group leaders should be listening to, because it shows the backlash is not really about the technology. It is about being handed change with no say in how it happens, no explanation of why, and no protection if it goes wrong. Companies that involve workers early, train them properly, and are honest about what AI will and will not change to their jobs are the ones that will avoid ending up in next year's negative numbers.