Can more tests prevent AI code from decaying?
They help, but they target the wrong layer. The common failure modes (duplicated logic, crossed architectural boundaries, orphaned code) all preserve behaviour, so behavioural tests pass. What is needed is a structural check about which parts may call which.
From Why AI-generated apps fall apart in week three, published 9 August 2026. That article is where the reasoning behind this answer is set out, including what it does not cover.