The e-commerce group had posted team-wide statistics on the use of AI by its employees, but recently limited access so that only employees and managers could see its statistics. Managers are discouraged from using the token to measure performance, according to a person familiar with the matter.
Meta employees have also engaged in so-called “tokenmaxing” to improve their position on internal leader boards.
The MeshClaw tool that some employees have used to boost their stats was inspired by OpenClaw, which became a viral sensation in February. OpenClaw allows users to run the agent locally on their own hardware, including computers and laptops.
Amazon’s MeshClaw can initiate code deployments, triage emails and interact with apps like Slack, according to people familiar with the matter.
The company said in a statement that the tool had “enabled thousands of Amazonians to automate repetitive tasks every day” and was an example of the group “empowering teams” to experiment and adopt AI tools.
“We are committed to the safe, secure, and responsible development and deployment of generative AI for our customers,” it added.
According to internal documents, more than three dozen Amazon employees worked on the in-house tool. A recent memo describing the bot states: “It dreams all night long to consolidate what it’s learned, keeps track of your deployments while you’re in meetings, and triages your emails before you wake up.”
Several Amazon employees said they were concerned about the security risks of AI tools that were allowed to act on a user’s behalf. This creates the risk of situations where the agent may make mistakes or perform unexpected actions.
One employee said, “The default security position scares me.” “I’m not just going to let it go and just do my job.”
© 2026 The Financial Times Ltd. All rights reserved. May not be redistributed, copied, or modified in any way.
<a href