Amazon’s cloud division suffered outages last year that were reportedly triggered by its own AI systems.
One disruption in December lasted 13 hours after an AI agent deleted and recreated part of its environment.
AWS said the incidents were limited and only one affected customer-facing services.
The company described the problem as user error caused by misconfigured access controls, not AI failure.
The outages have renewed concern about heavy reliance on a small number of cloud providers.
AWS runs large parts of the internet and holds major public-sector contracts.
The issue comes as Andy Jassy pushes efficiency measures and job cuts.
Amazon announced 16,000 layoffs in January after 14,000 the previous year.
Jassy said the reductions reflect cultural change, although he expects AI to shrink the workforce over time.
Some cybersecurity experts dispute Amazon’s explanation.
They argue AI systems can act faster than humans and may miss wider consequences.
Such tools often lack full context when making operational decisions.
Amazon says it has added new safeguards, including mandatory peer review for production access.
It also insists its AI tools act only within permissions set by human operators.
