YEREVAN (Azat TV) – Amazon Web Services (AWS) is firmly denying reports that its proprietary Kiro AI coding tool was responsible for a significant cloud outage in December 2025. The incident, which primarily impacted a single service within one of AWS’s Mainland China regions, has sparked a debate over the role of autonomous artificial intelligence in critical infrastructure following conflicting accounts from inside and outside the company.
Initially, a report by the Financial Times, citing sources familiar with the incident, alleged that AWS’s agentic Kiro AI coding tool autonomously decided to “delete and re-create the environment,” leading to a 13-hour service disruption. This report suggested broader impacts, including Amazon Alexa, Ring, and Prime Video services. However, AWS has since pushed back against this narrative, stating that the outage was a result of human missteps rather than AI autonomy.
AWS Denies Kiro AI as Outage Cause
In a statement to CRN, an AWS spokesperson clarified that the December 2025 outage was “the result of user error—specifically misconfigured access controls—not AI.” Amazon emphasized that the involvement of AI tools was merely a coincidence, asserting that “the same issue could occur with any developer tool or manual action.” The company further specified that the service interruption was an “extremely limited event” affecting only its AWS Cost Explorer service in one of its two regions in Mainland China.
AWS also highlighted that its Kiro AI tool is designed with inherent safeguards. By default, Kiro “requests authorization before taking any action,” ensuring that human engineers retain control and must configure which actions the AI can undertake. This contradicts the perception that Kiro operated entirely autonomously to cause the disruption.
Conflicting Reports and Employee Concerns
Despite AWS’s official stance, the Financial Times report, corroborated by several unnamed Amazon employees, painted a different picture. These sources suggested that the agentic nature of Kiro AI, which allows it to take autonomous actions on behalf of users, was indeed at the heart of the problem. Some senior AWS employees reportedly described these outages as “small but entirely foreseeable,” indicating a pattern rather than an isolated incident.
The December outage was reportedly not the first time AWS’s AI tools were linked to service disruptions. Multiple Amazon employees noted to the Financial Times that this was “at least” the second occasion in recent months, with another incident reportedly involving Amazon’s AI chatbot Q Developer. These internal perspectives underscore a growing concern among staff regarding the deployment and potential risks associated with increasingly autonomous AI tools within critical cloud environments.
Kiro AI Tool and Implemented Safeguards
Launched in July 2025, Kiro is an agentic coding service designed to assist developers by turning prompts into detailed specifications, working code, documents, and tests. It aims to automate tasks like generating documentation and unit tests, thereby streamlining development workflows. Amazon has actively encouraged its employees to adopt Kiro, reportedly setting an 80 percent weekly usage goal and closely monitoring adoption rates, while also selling access to the tool via a monthly subscription.
Following the December incident, AWS confirmed it has implemented “numerous additional safeguards,” including mandatory peer review for production access. These measures are intended to prevent similar occurrences and reinforce human oversight, regardless of the precise cause of the December outage. AWS also stressed that the event did not impact core services like compute, storage, database, or other AI technologies.
The ongoing dispute over the cause of the December AWS outage underscores a critical challenge facing the tech industry: establishing clear accountability and understanding the true extent of AI autonomy, especially as agentic tools become more integrated into essential infrastructure. The divergence between official corporate statements and internal employee accounts highlights the need for transparency as companies navigate the complexities of deploying advanced AI.