According to a Friday report by the Financial Times citing informed sources, Amazon's cloud business division experienced at least two service disruptions in December, both linked to errors in its internally developed AI tools.

The report stated that in mid-December, a customer-facing system at Amazon Web Services (AWS) suffered an outage lasting 13 hours. The cause was attributed to engineers allowing its Kiro AI programming tool to execute certain change operations.

The Financial Times reported that this intelligent tool, which can autonomously perform operations for users, independently decided to "delete and rebuild the operating environment."

An AWS spokesperson, responding via email to Reuters, described the outage as brief and attributed the cause to user error.

The spokesperson said, "This short-lived event was caused by user error—specifically, a misconfigured access permission, not an AI issue."

The spokesperson also stated that this service disruption was an "extremely limited scope" event, affecting only a single service within two regions in mainland China. It did not impact AWS's computing, storage, database, AI technology, or any other services.

In October of last year, Amazon Web Services experienced a large-scale global outage that significantly affected Amazon's own services as well as applications such as Reddit, Roblox, and Snapchat.