Amazon’s cloud division suffered at least one outage last year after an internal AI agent made a critical change to its environment.
The disruption lasted about 13 hours and forced systems to be deleted and rebuilt, according to reports.
AWS said the incident was caused by misconfigured access controls and described it as user error, not an AI failure.
The company added that only one event affected customer-facing services and that the impact was limited.
The outages have intensified debate about Amazon’s growing reliance on artificial intelligence.
Andy Jassy has said AI will improve efficiency and reduce routine work.
The company has already announced tens of thousands of job cuts.
Security experts argue AI tools can act faster than humans and may lack full operational context.
They warn that autonomous systems can make unexpected decisions, increasing risk if safeguards are weak.
Amazon said it has introduced stricter controls, including mandatory peer review for production access.
It insists AI does not cause more errors than human engineers and remains central to its long-term strategy.
