White‑collar workers resisting AI

Companies spent more on AI in 2026, and a lot of office workers responded by going around it. WalkMe said on April 9 that 54% of workers had bypassed company AI tools and done the task manually in the previous 30 days, while another 33% had not used AI at all. (walkme.com) That adds up to the “about 80%” figure in the coverage, but the split matters. One group tried the tool and chose the old way, and another group never touched it. (walkme.com) (finance.yahoo.com) The survey was not a tiny office poll. WalkMe said it covered 3,750 executives and employees in 14 countries, all at companies with at least 1,000 employees. (walkme.com) The surprise is that this comes after a year when workers were sneaking AI in through the side door. Coverage citing an earlier Massachusetts Institute of Technology study said employees at more than 90% of companies were using personal chatbot accounts for daily work even though only 40% of those companies had official large language model subscriptions. (finance.yahoo.com) So the story changed from “workers want AI before management does” to “workers do not trust the version management bought.” WalkMe found only 9% of workers trust AI for complex, business-critical decisions, while 61% of executives say they do. (walkme.com) Workers and bosses also disagree on whether the tools are any good. WalkMe found 88% of executives think employees have adequate AI tools, but only 21% of workers agree. (walkme.com) That gap gets expensive fast. WalkMe said average digital transformation budgets rose 38% year over year to $54.2 million, yet 40% of that spending was underperforming because people were not adopting the tools. (finance.yahoo.com) There is also a job fear problem sitting underneath the trust problem. Writer and Workplace Intelligence reported in March 2025 that 42% of executives said generative artificial intelligence adoption was “tearing their company apart,” and 31% of employees said they had sabotaged their company’s AI strategy, including by refusing to use the tools. (workplaceintelligence.com) Some of the resistance is about quality, not just fear. Bloomberg Opinion pointed to Princeton research published in February 2026 showing that while model capability improved, consistency stayed weak, with some systems scoring between 30% and 75% on reliability under identical conditions. (bloomberg.com) That helps explain why a worker might use artificial intelligence for a first draft but refuse to trust it on a contract, a forecast, or a customer email. A tool that is brilliant one time and wrong the next feels less like a calculator and more like an intern who occasionally invents facts. (bloomberg.com) The fight inside big companies is no longer mainly about getting access to AI. It is about whether managers can prove that the tool is reliable enough, useful enough, and safe enough that workers will use it without being pushed. (walkme.com) (workplaceintelligence.com)