AI Writes Code: Trust Lags Behind Rapid Adoption

A recent survey indicates that a majority of C++ developers are increasingly integrating AI into their daily coding practices. Approximately 58% of these professionals use AI for writing code and tests either daily, often, or sometimes. This rapid adoption, however, is not matched by a high level of trust, with a striking 78% of developers expressing concern that AI-generated content could be incorrect.

Despite the prevalent trust issues, popular AI tools like GitHub Copilot and ChatGPT are seeing widespread use. While developers find AI beneficial for understanding unfamiliar or legacy code, they express reservations about its utility in large, complex C++ projects, safety-critical code, and architectural decisions. This suggests AI is viewed as a supportive tool rather than a replacement for human programmers, especially given the potential for errors in sensitive applications.