Amazon Web Services (AWS) experienced at least two outages in December that were reportedly caused by errors involving its own AI tools.

First reported by The Financial Times, people familiar with the matter have linked the AWS outages to these tools.

The most significant incident lasted roughly 13 hours in the middle of December last year and took place after engineers reportedly allowed Kiro, an AI coding assistant, to make changes to a system used by customers. As a result of this decision, the tool went on to delete and rebuild the affected environment, which resulted in the disruption, according to reports.

People familiar with the matter reportedly said the agentic tool, which is able to take autonomous actions on behalf of users, had determined this was the best course of action.

However, a spokesperson from AWS told Capacity this incident was caused by misconfigured access controls, not AI behaviour.

The statement seen by Capacity reads: “This brief event was the result of user error—specifically misconfigured access controls—not AI. The service interruption was an extremely limited event last year when a single service (AWS Cost Explorer—which helps customers visualise, understand, and manage AWS costs and usage over time) in one of our two Regions in Mainland China was affected.

“This event didn’t impact compute, storage, database, AI technologies, or any other of the hundreds of services that we run. Following these events, we implemented numerous additional safeguards, including mandatory peer review for production access. Kiro puts developers in control—users need to configure which actions Kiro can take, and by default, Kiro requests authorisation before taking any action.” 

The December incidents followed a more serious outage in October, when a significant AWS failure disrupted the company’s own services and platforms, including Reddit, Snapchat and some essential services like Lloyds Bank and Venmo.

This incident also followed other significant outages last year that impacted companies like Colt Technology Services, Cloudflare and Microsoft Azure.

At the time, Amazon said the significant outage was caused by an issue in US-EAST-1, its largest cluster of data centres that powers the internet. Critical processes in the region’s database fell out of sync, causing a bug to occur and cause overwhelming delays. Much of these processes are automated and typically occur without human involvement.

Incidents like this contribute to current conversations over agentic AI, which are systems that take actions with minimal human input. As technology giants look to invest in them and further integrate them into critical infrastructure, there is debate over how these advanced systems could impact society.

In this vein, AI agents are becoming even more popular, with companies like Anthropic developing new models like Claude Opus 4.6 and OpenAI launching the Codex app to manage multiple agents at once.

“Given an AI agent full access to your computer and your web browser with all your sessions leads to incredible stuff – and that seems here to stay,” Atlman said at the Cisco AI Summit on Tuesday. “Companies that are not set up to quickly adopt AI workers will be at a huge disadvantage. And it’s going to take a lot of work and some risk.”

Related stories

AI agents are starting to talk: Are the networks ready?

Salesforce layoffs: The quiet job cuts behind a very loud AI boom

Anthropic forges Infosys AI deal amid widening tech layoffs

Amazon confirms new round of job cuts after internal email leak