Generative AI software development boosts productivity — and risk

This report examines how widespread adoption of AI coding assistants (e.g., GitHub Copilot) accelerates code creation and pull requests while increasing security risks, including more repositories with PII/payment data, exposed sensitive API endpoints, and a tenfold rise in APIs lacking authorization and input validation. Experts warn that AI often generates functional but insecure code and that developer over-trust plus insufficient review magnify risk; they recommend stricter governance, mandatory security reviews of AI-generated code, automated API security tests, and targeted developer training. The analysis also notes growing code churn and duplication and cautions that vulnerability backlogs and attack surface may expand without improved oversight.