A survey conducted by Fastly has found that nearly one-third of senior developers report that over half their shipped code is generated by AI tools, a significantly higher rate than among junior developers. The findings are based on a survey of 791 professional coders, with the results highlighting both the potential and challenges of AI integration in software development.
According to Fastly, 32% of senior developers—those with 10+ years of experience—indicated that more than half of their delivered code is AI-generated, a rate that is over two-and-a-half times higher than that reported by junior developers, who averaged 13%. This discrepancy may stem from the fact that more experienced developers are better able to assess the reliability and efficiency of AI-generated code, leading to a higher adoption rate among seniors. However, the use of AI tools is not without its challenges.
Senior developers are more likely to report that they invest time in fixing AI-generated code, with just under 30% saying they edit AI outputs enough to offset most of the time savings, compared to 17% of junior developers. This editing process is necessary because AI-generated code often requires manual adjustments to ensure it meets quality standards and is efficient. Despite this, 59% of senior developers say AI tools help them ship faster overall, compared to 49% of juniors. This suggests that the time saved by AI tools is often worth the effort of manual edits, especially for more experienced developers.
The survey also found that AI tools can boost developer job satisfaction. Nearly 80% of developers say AI makes coding more enjoyable, and this morale boost may have significant implications for the industry, particularly in a field grappling with burnout and backlogs. However, the study also highlights potential issues with AI productivity gains. A recent randomized controlled trial (RCT) of experienced open-source developers found that using AI tools actually led to a 19% increase in task completion time. This disconnect may be due to psychological factors, as the early speed gains from AI are often followed by cycles of editing, testing, and reworking that consume any time savings.
Fastly’s findings are echoed by other studies, with developers reporting that while AI tools can streamline the coding process, they often require a significant amount of manual intervention to maintain quality and efficiency. One developer noted that while AI saves time by using boilerplate code, it also requires manual fixes for inefficiencies, which can keep productivity in check. Additionally, the practice of green coding—considering energy use in code—has seen a sharp increase with experience, with 56% of junior developers reporting they consider energy use in their work, compared to 80% among mid- and senior-level engineers.
These findings illustrate the complex balance developers must strike between leveraging AI for productivity and ensuring code quality and efficiency. As AI continues to evolve, its role in software development will likely become even more prominent, with ongoing discussions about its impact on both the technical and human aspects of the field.