Amazon Web Services suffered at least two outages in December linked to its own artificial intelligence tools, the Financial Times reported, citing people familiar with the matter.
The more significant incident, lasting around 13 hours in mid-December, occurred after engineers allowed Kiro, an AI coding assistant capable of taking autonomous actions, to make changes to a system used by customers. The tool subsequently deleted and rebuilt the affected environment, triggering the disruption.
An AWS spokesperson pushed back on the characterisation, telling Reuters the incident was caused by misconfigured access controls rather than AI behaviour, and said the impact was confined to a single service in one region of mainland China, with no effect on compute, storage, database or other AWS services.
The December incidents follow a more serious outage in October, when a widespread AWS failure disrupted the company's own services as well as platforms including Reddit, Roblox and Snapchat.
The latest episode will nonetheless sharpen scrutiny of so-called agentic AI systems, which can act independently on users' behalf, as cloud providers move to integrate them more deeply into critical infrastructure.