× News Alerts AI News CyberSec News Let's Talk Local AI Bank Tech News Cyber Advisories Contact

Many Developers Report that AI Now Writes Most of Their Code

A new survey reveals that 42% of developers using AI state that at least half of their codebase is AI-generated. This growing reliance raises concerns about security and quality, as many bypass critical code review processes, highlighting a significant shift in software development practices.

Many Developers Report that AI Now Writes Most of Their Code

A significant shift is underway in the software development world, as a recent survey reveals that 42% of developers using artificial intelligence (AI) tools report that at least half of their codebase is now AI-generated. This finding, from Cloudsmith's Artifact Management Report 2025, underscores the rapid adoption of generative AI into daily programming workflows.

The survey, which polled 307 participants who all use AI daily, brings a concerning trend to light: despite the heavy reliance on AI-generated code, security practices are not keeping pace. Only 67% of respondents said they review AI-generated code before every deployment. This is particularly alarming given that 79.2% believe AI will exacerbate open-source malware threats. Furthermore, 41% considered code generation the greatest area of risk when it comes to AI-generated input in the software development cycle.

The adoption of AI is driven by productivity benefits. AI aims to automate repetitive tasks, allowing developers to focus on complex problem-solving. Tools like GitHub Copilot have shown they can speed up tasks. However, this speed can come at a cost. A study from GitClear uncovered concerning trends in code quality, noting that "code churn," the percentage of code that gets discarded shortly after being written, is projected to double in 2024.

Trust remains a critical issue. A 2025 Stack Overflow survey found that 84% of developers are using or plan to use AI tools, yet nearly half (46%) don't trust the accuracy of the output. This distrust has grown from 31% in the prior year. Interestingly, only 3.1% of developers state they “highly trust” AI-generated code.

Balancing speed with quality and security is the key challenge. Experts suggest that DevOps teams must thoughtfully integrate these technologies, establishing clear quality guidelines and stronger automated testing requirements. Human oversight remains irreplaceable for evaluating the broader impacts of AI suggestions. As AI continues to evolve, the software industry must develop robust practices to harness its power without undermining the stability and security of its foundations.

Subscribe for AI & Cybersecurity news and insights