Amazon corporate employees are pushing back against the company's internal mandate to adopt AI tools, telling The Guardian that the push is counterproductive rather than helpful. Workers describe the tools as "half-baked" and error-prone, saying the time required to verify AI outputs and correct mistakes often exceeds any gains. "I and many of my colleagues don't feel that it actually makes us that much faster," one software developer told The Guardian, even as management messaging insists AI must accelerate pace and that "speed is the number one priority." Amazon spokesperson Montana MacLachlan disputed the framing, telling Gizmodo that the Guardian report drew "broad conclusions from a small number of interviews" and that employees are spending less time on rote work and doing more creative, impactful work as a result.
The Amazon complaints are backed by data from a large-scale study conducted by workforce analytics firm ActivTrak, which analyzed activity across 163,638 employees at 1,111 organizations over three years. Across every measured work category, AI adoption increased workloads. Email volume rose 104%, chat and messaging surged 145%, and time spent with business management tools climbed 94%. "AI is being used as an additional productivity layer, not a substitute for existing work," the researchers concluded. Even in cases where AI did speed up individual tasks and free up time, that time was simply filled with additional work — boosting corporate output metrics while doing nothing to ease the burden on individual employees.
Hacker News commenters responding to the story described being required to demonstrate active use of tools like Microsoft Copilot to satisfy KPI metrics, regardless of actual utility, and summarized the dynamic bluntly: becoming more productive simply results in being assigned more work. Former Google executive Mo Gawdat, quoted in the Gizmodo piece by reporter Ece Yildirim, drew parallels to prior technology cycles — social media promised connection and delivered loneliness; mobile phones promised work-life balance and delivered always-on availability. His argument is that technology amplifies existing values, and in a profit-driven economy, that means amplifying output demands rather than worker leisure.
The ActivTrak data punctures the standard <a href="/news/2026-03-14-perplexity-launches-personal-computer-ai-agent-platform-for-enterprise">enterprise AI pitch</a> — that automation frees workers from drudgery. What it actually shows, at scale, is that freed-up time gets immediately reallocated by management. Without structural changes to how productivity gains are distributed, deploying more agents into workflows just means more work gets done by the same number of people. The company captures the upside; the worker absorbs the load.