In December, Amazon experienced two minor outages within its cloud infrastructure due to its Kiro AI system executing tasks that resulted in a 13-hour disruption. This incident, reported by Financial Times, is less severe than the significant AWS outage from October, which was attributed to user error, not directly to AI. An AWS spokesperson clarified that the December outages did not affect major services, and the blame lies with the engineers who allowed AI interventions. Despite the concerns surrounding AI’s role in operational reliability, the recent outages did not lead to catastrophic consequences like previous high-profile disruptions across platforms such as YouTube, Verizon, and Microsoft 365. As companies increasingly depend on a few cloud providers, the risk of widespread outages remains. The implications underscore the importance of careful AI integration in critical infrastructure, as well as ongoing discussions about the frequency and impact of internet outages overall.
Source link
