Amazon’s cloud division reportedly suffered at least one service disruption last year after an internal AI agent made a critical change to its environment.
The tool autonomously deleted and recreated a system component, causing a lengthy interruption, according to the Financial Times.
AWS underpins large parts of the internet, so outages raise concerns about concentration of digital infrastructure.
The company said the incident was limited in scope and described it as user error caused by misconfigured access controls.
It added that core services such as computing, storage and databases were not affected.
The report comes as Andy Jassy continues major job cuts at Amazon.
About 30,000 roles have been eliminated since late 2024.
Jassy has said AI will improve efficiency but denied that layoffs directly replace workers.
Some cybersecurity experts question Amazon’s explanation.
They argue AI systems can act faster than humans and may not grasp the wider impact of their actions.
Their limited context can increase the risk of unexpected failures.
Amazon says new safeguards are in place, including mandatory peer review for production access.
The company maintains that AI tools remain under human control and do not inherently cause more errors.
