What are the security risks associated with AI-generated code?
Veracode's 2025 GenAI Code Security Report evaluated more than 100 LLMs across 80 coding tasks and found that 45% of tasks introduced at least one known security vulnerability. In Java, failure rates exceeded 70%.
Answered in
The 7.7% Mirage: Why AI Code Generation Is Multiplying Maintenance, Not VelocityAI writes code many times faster, yet median PR throughput rose just 7.76% while code churn doubled. The bottleneck moved downstream.
Read the full analysisOther questions this article answers
More development best practices questions
- How much faster does AI help developers write code?
- What is code churn and how has AI impacted it?
- Why did developer PR throughput increase by only 7.76% despite a 65% increase in AI tool usage?
- How does AI impact the Total Cost of Ownership (TCO) of software?
- Does a larger context window prevent agent loops?
- Should every failed test trigger git reset --hard?
- Do agent applications require a modular monolith?
- Can one database transaction roll back an entire agent workflow?
Every answer on Crashtech is written by the editor of the article it comes from — never auto-summarised. Browse all answers or the Development Best Practices beat.