Amazon Employees Face Pressure to Boost AI Usage—But Some Are Gaming the System

Amazon employees say the company is increasingly pressuring them to integrate artificial intelligence (AI) into their daily workflows. However, the specifics of how AI should be used remain vague, creating an environment where workers may waste resources on extraneous tasks to meet rising expectations.

According to a new report by The Financial Times, some Amazon employees are reportedly using the company’s internal AI tool, MeshClaw, to generate unnecessary AI agents—not to enhance productivity, but to artificially inflate AI usage metrics. The employees claim Amazon is monitoring their consumption of AI tokens, incentivizing some workers to prioritize quantity over quality when adopting the technology.

Employees Describe Rising Pressure and Perverse Incentives

Several anonymous Amazon employees spoke to The Financial Times about the growing pressure to use AI tools, which they say is negatively impacting their workplace.

“There is just so much pressure to use these tools,” one Amazon worker said. “Some people are just using MeshClaw to maximize their token usage.”

While Amazon has reportedly assured employees that AI usage statistics would not factor into performance evaluations, not all workers are convinced.

“Managers are looking at it,” another employee said. “When they track usage, it creates perverse incentives, and some people are very competitive about it.”

The interviewed employees claim Amazon has set a target for 80% of developers to use AI each week, with token consumption tracked on an internal leaderboard. However, an Amazon representative denied the existence of any company-wide metric or leaderboard, stating that employees can only view their own AI usage on personal dashboards.

MeshClaw: A Powerful but Controversial Tool

MeshClaw, the tool some employees are using to inflate AI usage, draws inspiration from OpenClaw, another AI tool known for both its productivity potential and risks. Unlike many AI models, OpenClaw and MeshClaw operate locally on users’ hardware, granting them significant autonomy.

Earlier this year, the director of alignment at Meta Superintelligence Labs went viral when OpenClaw nearly deleted her entire email inbox, highlighting the dangers of granting AI excessive access.

At Amazon, MeshClaw can be used to deploy code, sort through emails, and interact with apps like Slack. A recent internal memo described MeshClaw’s capabilities in ambitious terms:

“MeshClaw dreams overnight to consolidate what it learned, monitors your deployments while you’re in meetings, and triages your email before you wake up.”

Not all employees are comfortable with this level of autonomy.

“The default security posture terrifies me,” one employee said. “I’m not about to let it go off and just do its own thing.”

Amazon Defends AI Adoption Strategy

In response to inquiries, an Amazon spokesperson told Fast Company that MeshClaw “was built by a small team and it enables thousands of Amazonians to automate repetitive tasks each day, freeing up time for employees to be more strategic and solve bigger customer problems.”

The spokesperson added, “This is just one example of how we’re empowering teams to experiment with AI and how we’re supporting employees’ adoption of AI.”