A recent survey shows that more than half of the code shipped by senior developers is now generated by AI tools. This is a surprising shift, especially since many experts worry about the security risks AI could introduce into software. The survey, conducted by cloud platform provider Fastly, included 791 professional developers and was carried out in mid-July 2025.

Senior Developers Rely Heavily on AI for Coding

According to the survey, 32% of senior developers—those with over 10 years of experience—said that more than half of their code was created with help from AI. In comparison, only 13% of junior developers, with fewer years of experience, reported the same level of AI use. This suggests that more experienced programmers are not only using AI more often but also trusting it more when building production-ready code.

Fastly’s findings raise interesting questions about the roles of experience and trust in AI. Despite concerns about AI-generated code potentially having vulnerabilities, many developers seem to find AI tools helpful for speeding up their work. Nearly one in three developers admitted they often had to fix or edit AI-generated code, which sometimes took up most of the time saved. Still, a majority felt that AI made them faster overall.

The Benefits and Challenges of Using AI in Coding

Tools like GitHub Copilot, Google’s Gemini, and Claude are popular among developers. Fastly’s survey indicates that about 59% of senior developers said these AI tools helped them ship code faster, compared to 49% of junior developers. Interestingly, more than half of junior developers said AI made them moderately faster, but only 39% of senior developers agreed. Still, senior developers were twice as likely to report significant speed improvements.

One reason might be that senior developers are better at catching mistakes made by AI and fixing them quickly. Nearly 80% of all respondents said AI made coding more enjoyable. For some, it’s the chance to skip repetitive tasks, while others get a thrill from seeing code appear on demand.

However, using AI isn’t always seamless. Some developers find that AI sometimes makes assumptions that don’t match their intentions, forcing them to redo parts of the code. A junior developer mentioned having to go back and redo AI-generated work when it didn’t align with their goals. Meanwhile, senior developers tend to spend more time fixing AI output, but they still see clear benefits in speed and productivity.

What This Means for Developers and the Future of Coding

Overall, the survey shows that AI is becoming a major part of software development, especially among more experienced programmers. They seem to trust and leverage AI tools more than their less experienced counterparts. This trend could reshape how teams work, with AI handling more of the routine coding tasks and humans focusing on complex problem solving.

Fastly’s findings also suggest that AI can make development more enjoyable and efficient, even if it sometimes requires extra effort to fix AI mistakes. As AI tools continue to improve, they might become an essential part of every developer’s toolkit, helping teams deliver better software faster.

In the end, AI seems poised to change the landscape of software creation. Developers who learn to work alongside these tools will likely have an edge in the fast-evolving tech world.

Inspired by

• https://www.infoworld.com/article/4049949/senior-developers-let-ai-do-more-of-the-coding-survey.html

Sources

• gmpg.org
• fastly.com
• podigee.com
• podigee.com