SmartBear: 81% of leaders trust AI to catch its own errors, evidence lags

SmartBear's State of Software Quality and Testing 2026 report found 81% of organizational leaders believe AI reliably catches its own errors, while only 64% of practitioners trust it. 46% of teams shipped AI-generated code that later failed, yet 69% of those still reported high confidence in it.

• 81% of leaders trust AI self-validation vs 64% of developers
• 46% of teams shipped AI-generated code that later failed
• 92% say AI is the primary tester of its own code
• 84% use at least one form of human review for AI tests