Cybersecurity Risks and Software Vulnerabilities in AI-Generated Code

The increasing adoption of Artificial Intelligence (AI)–based code generation tools has introduced new challenges for software security. This exploratory study analyzes 120 AI-generated code samples to identify recurring security vulnerabilities across common programming tasks. Code samples were generated using a consistent prompting strategy and manually reviewed using the OWASP Top 10 security framework.

The findings indicate that a substantial portion of the analyzed samples contained identifiable security weaknesses, particularly in areas related to authentication, database operations, and input handling. However, these observations reflect potential vulnerability patterns rather than definitive generalizations. The results are intended to contribute preliminary empirical insights into the security characteristics of AI-generated code and to inform future large-scale and comparative research.