Amazon Scraps AI Usage Leaderboard After Engineers Game the System 2026

Amazon has scrapped an internal experiment that measured employee performance based on their use of corporate AI tools, after engineers began gaming the system to climb a leaderboard. The initiative, which aimed to boost AI adoption, instead led to wasted computing resources and meaningless tasks.

What Happened: Amazon's AI Usage Experiment

Amazon implemented a tool to track how often employees used its AI-powered coding assistant, Kiro. The goal was ambitious: get 80% of its programmers to use AI every week. But within weeks, the company noticed something troubling. Engineers were using AI for absurd, low-value tasks just to increase their usage metrics.

The experiment was canceled entirely after it became clear that the approach was backfiring. Instead of driving meaningful AI adoption, it encouraged employees to game the system, inflating costs and undermining productivity.

The Kirorank Leaderboard: A Gamification Gone Wrong

The service at the center of the controversy was an internal scoreboard called Kirorank. It tracked developers' activity within Kiro, Amazon's version of Claude Code, and ranked them based on how much they used the AI tool.

To climb the ranks, developers deployed autonomous agents based on MeshClaw—Amazon's version of OpenClaw—to run processes in loops and consume AI resources without producing any real value. This not only wasted computational power but also drove up infrastructure costs significantly.

Why Forced AI Adoption Backfires

Amazon's experience highlights a common pitfall: forcing AI adoption through metrics can lead to unintended consequences. When employees are evaluated on usage rather than outcomes, they may prioritize looking busy over being effective.

This is not the first time gamification has led to gaming. Similar issues have arisen with other productivity metrics, from lines of code to hours logged. The lesson is clear: measure what matters, not just activity. The background to this story is set out in Amazon Makes Alexa+ Free on Fire TV to Boost AI Adoption. A related thread runs through AI Adoption in Offices: Why Employees Resist and What It Means. Earlier coverage in Microsoft AI Diffusion Report: Global AI Adoption Map Revealed shows how the argument developed.

Key Takeaways for Business Leaders

• Metrics drive behavior: Employees will optimize for whatever is measured, even if it's counterproductive.
• AI adoption should be organic: Forcing usage can lead to waste and resentment.
• Focus on outcomes: Measure the value AI brings, not just how often it's used.
• Watch for unintended costs: Gaming the system can inflate cloud and infrastructure expenses.
• Learn from failures: Amazon's quick cancellation shows the importance of adapting when experiments go wrong.

What This Means for the Future of AI in the Workplace

Amazon's decision to abandon the leaderboard doesn't mean the company is giving up on AI. Instead, it's a recognition that top-down mandates can backfire. As AI tools become more integrated into workflows, companies need to find better ways to encourage adoption without encouraging gaming. A closer look at the same subject appears in What Is the Largest Company in the World? It Depends on the Metric. Anyone following the wider picture can start with Amazon Launches Fast-Track GLP-1 Weight Loss Program. The same theme surfaces in Ai Video: latest articles, which takes a different angle.

The incident also raises questions about how AI usage should be evaluated. Instead of raw usage numbers, firms might look at code quality, project outcomes, or time saved. The focus should be on value, not volume. Context on how this played out before sits in Amazon Employees Resist AI Tools With Memes and Skepticism. A companion piece, AI at Work Saves Time Then Adds Hours of Fixes, approaches the same subject from another angle.

Comparison of AI Adoption Metrics

FAQ