Flaky tests won't disappear just because AI writes them — flakiness comes from environment dependencies, timing issues, and shared state, not from how the test code was written. But AI can dramatically reduce the maintenance burden of flaky suites by auto-diagnosing and rewriting tests that fail intermittently. The real win: AI that watches your CI pipeline, identifies flaky patterns, and proposes targeted fixes. That's where generative AI in testing actually delivers — not replacing test writers but becoming the best debugging partner for test infrastructure.