What happened
Amazon called a large group of engineers to a meeting on Tuesday to examine a series of service disruptions the company associates with artificial intelligence-related activity. The meeting was convened to review recent events and internal findings, according to a briefing note prepared for the session.
What the briefing identified
The briefing note cited a "trend of incidents" occurring over recent months and identified several contributing characteristics. Among those listed were a "high blast radius" for incidents and the presence of Gen-AI assisted changes to systems. The note also singled out "novel GenAI usage for which best practices and safeguards are not yet fully established" as a factor complicating operations.
Recent outages and operational effects
This month Amazon's website and shopping application were unavailable for nearly six hours in an episode the company attributed to an incorrect software code deployment. During that outage, customers were unable to complete transactions or access functions such as checking account details and viewing product prices.
Incidents linked to internal AI tools
Amazon Web Services has reported at least two incidents connected to the use of AI coding assistants that the company has been deploying internally. Those incidents form part of the pattern the company is studying and were cited in the meeting briefing as examples of where Gen-AI interactions with engineering workflows have had operational consequences.
Context within the company
The meeting brings together engineers to assess the recurring issues identified in the briefing note and to consider whether changes to procedures, deployment practices or safeguards are needed as the company continues to adopt generative AI tools across development teams. The briefing language highlights uncertainty around established practices for newer GenAI use cases and the potential for wide-ranging impact when incidents occur.