Security and Quality Issues Are Rising in AI-Generated Code
While the share of AI in the software processes of tech giants is growing rapidly, new research reveals that this code contains more bugs, maintenance issues, and security vulnerabilities.
While the proportion of code written by artificial intelligence at major companies like Google and Microsoft is rising rapidly, recent research shows that this code contains significantly more bugs and security vulnerabilities compared to human-written code.
AI Ratios in the Software World
Google announced in late 2024 that approximately 25% of its new code was generated by artificial intelligence. This rate rose to 50% a year later and reached 75% in April 2026. Microsoft CTO Kevin Scott predicts that 95% of software code will be generated by artificial intelligence in the coming years.
Research Findings and Error Rates
Recent research by CodeRabbit revealed that AI-generated code produces an average of 10.83 issues per request. In human-written code, this figure was measured at 6.45, determining that AI code produces 1.7 times more issues.
Maintenance Challenges and Security Vulnerabilities
Findings published by Sonar showed that over 90% of issues in AI models are code smells that create long-term maintenance headaches. ProjectDiscovery emphasized that there is a growing gap between the speed of AI code generation and the capacity of security teams.