Sonar's State of Code Developer Survey, based on responses from more than 1,100 professional developers, found that AI now accounts for roughly 42% of code committed to production and is on track to reach 65% by 2027, but the trust picture behind that adoption curve is far shakier than the growth numbers suggest: 96% of developers say they do not fully trust AI-generated code, yet only 48% report actually verifying it before it ships. That gap between stated distrust and actual verification behavior is the report's most important finding, because it means a large share of AI-generated code is reaching production on faith rather than review, precisely because reviewing it properly has become its own bottleneck. The survey found developers now spend 11.4 hours per week reviewing AI-generated code versus 9.8 hours writing new code themselves, a reversal of the pattern from just two years earlier when writing code consumed more time than reviewing it. 38% of developers specifically said reviewing AI-generated code requires more effort than reviewing a human colleague's code, which tracks with what a lot of engineering teams have anecdotally reported: AI-generated code tends to be locally plausible and syntactically clean while sometimes missing broader context about why a codebase does things a certain way, which makes surface-level review misleading and genuinely thorough review slower, not faster. For engineering leaders, the practical implication is that raising AI code-generation adoption without a corresponding investment in review tooling and review time budgets is quietly building a verification debt, and teams should treat "percentage of code that is AI-generated" as a much less useful metric than "percentage of AI-generated code that has actually been properly reviewed."