We scanned 1,868 AI-built apps for production readiness, and audited our scanner (pathtoship.com)

A recent audit of 1,868 public apps built with AI coding tools reveals significant insights into their production readiness and highlights the importance of transparency in automated scanning technologies. The analysis from PathToShip showed that only 23% of the apps met the 80-point production-ready threshold, with a mean score of 68.3. While scalability (mean score of 80) and cost efficiency (mean score of 88) were strong points, production readiness (mean score of 56) and security (mean score of 65) showed considerable gaps. Notably, 24% of the apps had at least one critical finding, which raises concerns about the reliability of AI-generated code in real-world applications. The significance of this study lies in its rigorous approach to auditing its scanning method, which initially had a 42% false-positive rate for critical findings. After revisions, this rate was reduced to about 25%. The systematic checks conducted across seven dimensions, including security and architecture, emphasize that while AI tools can produce efficient code, they often fall short in deployment readiness, handling secrets securely, and ensuring overall robustness. This transparency in methodology not only enhances credibility but also invites further collaboration from the community to improve AI-driven coding practices.