Amazon Workers Mock Internal AI Tool 'Kiro' as Enterprise Adoption Struggles

A report from 404 Media has surfaced internal Amazon communications revealing widespread worker frustration with the company's push to adopt AI tools, particularly a coding assistant called Kiro. The findings parallel earlier reports of Google employee backlash against internal AI mandates, suggesting a broader pattern of resistance when enterprise AI deployments are driven by performance metrics rather than demonstrated utility.

Memes shared in Amazon's internal Slack channel, #actual-aws-memes, mock Kiro for confidently misrepresenting what it can do. One widely circulated example shows an iceberg with the tool saying "Confirmed, I have the full picture," while only addressing the surface ice. Workers nicknamed the tool "Sloppenheimer" — combining "slop" (a term for low-quality AI output) with a reference to the Oppenheimer film.

The frustration is compounded by a separate failure: Amazon recently canceled an AI leaderboard system designed to track and incentivize employee AI usage. Workers had inflated their scores with low-quality "junk prompts," resulting in decreased actual productivity. Amazon confirmed the shutdown to 404 Media.

Amazon responded to the coverage: "This handful of comments doesn't reflect what we hear from most Kiro users, we still appreciate the chance to learn from the feedback."

Key Takeaways

• Amazon workers are openly mocking internal AI tool "Kiro" in company Slack for inaccuracy and overconfidence
• Amazon shut down an AI usage leaderboard after employees gamed it with junk prompts, worsening actual output
• The backlash mirrors a similar pattern at Google, pointing to a systemic enterprise AI adoption challenge
• Performance-metric-driven AI mandates appear to surface tool weaknesses and erode worker trust

Read the full article on ExtremeTech