Meta launches AI leaderboard to track employees, then shuts it down within days

Meta briefly rolled out an internal leaderboard to track how employees use artificial intelligence tools, before shutting it down within days after concerns emerged over sensitive data being shared externally.

The system, designed to make AI usage central to day-to-day work, ranked employees based on how extensively they used generative AI tools. However, the initiative was withdrawn after internal metrics began circulating outside the company, according to India Today and The Information.

A gamified push to drive AI adoption

The leaderboard, known internally as “Claudeonomics”, tracked employee activity by measuring “tokens” — the units that quantify how much input and output is processed by large language models.

According to The Information, the system monitored the top 250 employees, ranking them by total token consumption. It also introduced gamified elements, awarding badges such as “Token Legend”, “Model Connoisseur”, “Cache Wizard” and “Session Immortal”.

The name referenced Anthropic’s Claude model, which has reportedly been widely used within Meta, particularly for coding-related work.

Data cited in the report showed the scale of adoption. Employees collectively used around 60 trillion tokens over 30 days, with the highest individual usage reaching 281 billion tokens, underscoring how deeply AI tools are being integrated into workflows.

Pulled over data sharing concerns

Despite its internal positioning as an engagement tool, Meta moved quickly to dismantle the system.

“It was meant to be a fun way for people to look at tokens, but due to data from the dashboard being shared externally, we’ve made the decision to shutter Claudeonomics for now,” the company said in a message cited by Gizmodo.

Meta has not released updated figures on token usage following the removal of the leaderboard.

Wider industry shift towards measuring AI use

The episode reflects a broader trend in the technology sector, where companies are increasingly tracking how employees use AI tools and, in some cases, linking it to productivity.

This approach — sometimes referred to as “tokenmaxxing” — centres on maximising AI usage as a performance indicator.

Industry leaders have begun signalling support for such models. Nvidia chief executive Jensen Huang has suggested that engineers should make extensive use of AI resources, even proposing that companies allocate token budgets as part of compensation packages, according to remarks at the GPU Technology Conference 2026 and on the All-In Podcast.

Under such a framework, AI compute could sit alongside salary, bonuses and equity as a core component of pay.

Experiment ends, but questions remain

Meta’s short-lived experiment highlights both the growing importance of AI in workplace productivity and the risks tied to internal data visibility.

While the company has pulled the leaderboard for now, the underlying direction is clear: firms are moving to quantify and incentivise AI usage, even as governance and data controls struggle to keep pace.