Amazon faces an uncomfortable contradiction. The company intends to require additional human oversight when generative AI tools are used by developers, with junior and mid-level engineers needing approval from senior engineers before AI-assisted code changes can be released. Yet Amazon has laid off more than 30,000 employees since October 2025, leaving the company with far fewer people to perform that essential gate-keeping function.
The policy emerged from a series of production incidents. In December 2025, the Kiro AI coding tool autonomously deleted and recreated an AWS Cost Explorer environment, triggering a 13-hour outage in a China region. Several incidents involved software changes produced with the assistance of generative AI coding tools, and Amazon's internal messages noted the use of these tools remains new in many teams and that established safeguards are still incomplete. These failures prompted senior leadership to act.
The underlying issue is real. When AI can generate hundreds of lines of production code in minutes, traditional code review processes become inadequate overnight. AI-assisted code introduces a fundamental shift: developers now ship code they didn't fully author and may not completely understand. This knowledge gap exposes teams to risks that conventional code reviews, designed for human-authored work, may not catch.
Yet the company's path forward contains a logical flaw. Adding a mandatory senior engineer approval gate sounds reasonable on its face; someone with deeper system knowledge should validate changes before they go live. But the senior approval rule creates a bottleneck at the senior engineer level that will slow deployment velocity, the opposite of what Amazon wanted from AI tools. Senior engineers are scarce resources. Asking them to review every junior engineer's AI-assisted change could easily become a choke point.
Amazon's own framing of the problem suggests tension between competing priorities. The company promoted AI coding tools as a way to accelerate development across its enormous infrastructure footprint. AWS underpins countless businesses and services globally, and Amazon has invested billions into maintaining competitive speed and reliability. Yet the path to security now requires human judgment that the company has systematically reduced through its recent redundancy programme.
Whether this plays out as a temporary adjustment or a fundamental challenge will become clear in coming months. Amazon will likely implement stricter AI code review protocols, including mandatory senior approvals, to mitigate outage risks, but this could slow development velocity while improving reliability for e-commerce and AWS services. That trade-off may prove necessary. But executing it with 30,000 fewer people suggests the maths will remain awkward for some time.