Artificial intelligence is rapidly becoming a co-pilot for software engineers. But a series of outages at Amazon.com, Inc. (NASDAQ:AMZN) suggests the technology may still come with operational risks — a point that prompted Elon Musk to urge restraint.

Musk wrote “Proceed with caution” on X in response to a post discussing a report by Rafe Rosner-Uddin of the Financial Times, which said Amazon engineers are examining several incidents linked to AI-assisted coding changes.

The report said Amazon’s retail technology teams are holding a large engineering meeting to analyze a recent rise in outages and system disruptions.

AI Tools Meet Real-World Systems

Generative AI tools are increasingly being used to help developers write, edit and deploy software. While the tools promise faster coding cycles, they can also introduce unexpected changes in complex production environments.

According to the FT report, Amazon’s internal review cited instances in which AI-generated code changes contributed to incidents that affected the availability of parts of the company’s online infrastructure.

In an internal email referenced by the FT, Amazon senior vice president Dave Treadwell told employees the reliability of the company’s retail platform “has not been good recently,” prompting a deeper internal review.

One outage earlier this month reportedly disrupted Amazon’s website and shopping app for several hours after a faulty software deployment prevented customers from completing transactions.

New Guardrails Around AI Coding

Amazon is now tightening oversight of AI-assisted development. The FT reported that engineers will increasingly require senior approval before implementing changes generated with AI tools.

The issue has also surfaced at Amazon Web Services, where an AI coding assistant previously disrupted an internal service by automatically modifying part of the system environment.

Amazon said operational reviews like these are routine as the company works to improve system reliability.

But the incidents highlight a broader challenge facing the technology industry: balancing the speed and efficiency of AI-generated code with the stability required to run large-scale digital platforms used by millions of customers.

Image: Shutterstock