Choose Wisely: AI-Generated Coding Risk Varies, A Lot

Secure Code Warrior's new AI Trust Index evaluated 1,760 codebases from 16 frontier AI models, finding an average of 15 vulnerabilities per codebase. No single model dominated — performance varied drastically by framework, making model-framework pairing the real risk factor. Organizations should match security gates and developer training to high-risk frameworks rather than simply picking the highest-ranked AI coding tool.