Amazon Web Services suffered a 13-hour system outage in December as a result of the actions of its AI coding assistant Kiro, according to financial Times. Several unnamed Amazon employees reported foot AI agent Kiro was responsible for an incident that affected AWS services in parts of mainland China in December. People familiar with the matter said the tool opted to “delete and recreate the environment” it was working on, which led to the outage.
While Kiro typically requires sign-off from two humans to push changes forward, the bot had permission from its operator, and a human error there allowed more access than expected.
Amazon described the December disruption as an “extremely limited incident,” compared to a larger outage in October that knocked out online services like Alexa, Fortnite, ChatGPIT and Amazon for hours. An outage in which no one is trapped in their smart bed is a lucky escape.
This isn’t the first time AI coding tools have caused problems for Amazon. A senior AWS employee said the December outage is the second production outage involving an AI tool in the past few months, the other being linked to a developer of Amazon’s AI chatbot Queue. The employee described the outage as “small but completely predictable.” Amazon said the second incident had no impact on “customer facing AWS services.”
Amazon blamed human error for the problems, not rogue bots, and said it had “implemented numerous security measures” such as employee training following the incident. The company said it was “a coincidence that AI tools were involved” and stressed that “the same problem could have occurred with any developer tools or manual action.” That’s true, and although I’m not an engineer, I think no one would knowingly scrap and rebuild something to make a difference under the most severe circumstances.
