Amazon's cloud 'hit by two outages caused by AI tools last year' (www.theguardian.com)

Amazon's cloud computing division, AWS, experienced two significant outages over the past year, attributed to its own AI tools. The most notable incident occurred in December, resulting in a 13-hour interruption when an AI agent autonomously deleted and recreated a segment of its environment. While AWS has faced several outages throughout the year, notably affecting major online services, Amazon downplayed the AI's role, claiming the incidents were primarily due to user error rather than AI malfunction. These outages raise critical questions about the reliability of AI in managing essential cloud infrastructure, particularly as Amazon has announced substantial layoffs, suggesting a growing reliance on AI for operational efficiency. Experts highlight the limitations of AI agents, which lack the contextual understanding that human engineers possess, potentially leading to significant errors. The incidents reflect broader concerns within the AI/ML community regarding the balance between automation and human oversight, especially as companies accelerate their integration of AI technologies while simultaneously reducing their workforce.