Amazon pushes AI into every task, even as it slows productivity

Amazon’s AI rollout is now a company‑wide mandate, even as workers on the front lines report that the tools are doing more harm than good. Software engineer “Dina” in New York, who was hired two years ago to write code, says the internal system called Kiro “hallucinates” and spits out buggy snippets that she must spend hours untangling or discard entirely. “It feels like trying to AI my way out of a problem that AI caused,” she told The Guardian, adding that the speed boost promised by management is nowhere to be seen in her day‑to‑day workflow. Within days of the interview, Dina was laid off, underscoring the precariousness of a workforce being asked to adopt imperfect tools while headcount shrinks.

The experience is not isolated. A supply‑chain engineer with a decade at Amazon, identified only as Lisa, says AI assistance works roughly one‑third of the time. Even when the output is usable, she must “consult with colleagues to verify and correct” the results, a process that “takes up more time than if she’s done the task without AI,” according to The Guardian. Employees across roles—from user‑experience researchers to data analysts—describe a “haphazard” rollout that forces them to log usage while the company silently harvests the data to train future bots. The anonymity granted to these sources highlights a growing fear that the push to embed AI is as much about building a replacement workforce as it is about productivity gains.

Amazon’s leadership frames the initiative differently. Spokesperson Montana MacLachlan told The Guardian that “hundreds of thousands of corporate employees” are experimenting with AI and that “the vast majority of our teams are getting a lot of value out of the AI tools they use day‑to‑day.” Yet the same period has seen Amazon cut roughly 30,000 corporate jobs—about 10 % of its 350,000‑strong workforce—part of a broader wave of AI‑related layoffs that also hit Block, Pinterest and Autodesk. While Block’s CEO Jack Dorsey openly blamed AI for a 40 % staff reduction, Amazon has offered mixed messages, saying AI will eventually drive reductions but denying that the recent cuts were directly AI‑driven.

The company’s financial commitments to AI are massive. In February, Amazon announced a $200 billion spend on AI infrastructure for the year and a $50 billion partnership with OpenAI, signaling confidence that the technology will underpin future growth. Yet the internal friction suggests the rollout is outpacing the maturity of the tools themselves. Reuters reported that Amazon is simultaneously expanding its AI footprint into consumer‑facing services, launching a healthcare assistant on its website and app, a move that mirrors the same “AI‑first” philosophy being forced on its corporate staff. The contrast between high‑visibility consumer products and the day‑to‑day struggles of engineers hints at a strategic gamble: Amazon is betting that the long‑term payoff of ubiquitous AI will outweigh short‑term productivity losses.

Analysts note that the pressure to adopt AI at scale may be eroding morale as much as it is inflating expectations. Workers describe a culture where “you don’t look at the problem and go, ‘How do I use this hammer I have?’” but instead “ask whether it’s a problem for a hammer or something else,” a sentiment echoed by Lisa in The Guardian. The disconnect between corporate messaging—speed and efficiency as “the number one priority”—and the lived reality of buggy code and extra verification steps raises questions about the true cost of Amazon’s AI ambition. If the company cannot reconcile the promise of AI‑driven acceleration with the on‑the‑ground experience of its employees, the push may end up costing more in lost productivity and talent than it saves in future automation.