Dive Brief:
- Developers are using AI more often but aren’t becoming bigger fans of the technology in the process, according to a recent Stack Overflow survey. The public question-and-answer platform for tech pros surveyed more than 49,000 developers in the 15th year of the report.
- Positive sentiment for AI tools dropped 10 percentage points this year. More developers said they actively distrust AI tools than those who said they trust them, with experienced developers being the most skeptical.
- The biggest frustration among developers was dealing with AI tools that are “almost right, but not quite,” according to the report. The time it takes to debug AI-generated code was another prickly pain point for respondents. While AI use is up, developers are most resistant to applying the technology for high-responsibility tasks, such as deployment, monitoring and project planning.
Dive Insight:
Integrating AI into developer workflows has brought changes to the job, and analysts expect more shifts ahead.
Engineers and developers are aware of potential AI-driven job impacts, from calls for massive upskilling to an upheaval of existing techniques.
“The hard part of adopting new tools always is that people are just so used to their environment,” Ryan Ries, chief AI and data scientist at managed services and consulting provider Mission Cloud, told CIO Dive earlier this summer. “Now, it’s like ‘I’ve got to learn everything over again,’ and it’s super frustrating.”
For tools to become fully integrated into workflows, developers have to believe the AI add-ons are valuable and can complete the task at hand, Ries said.
Despite hesitation elsewhere, developers are most likely to use AI to research, generate synthetic data and learn new concepts, according to the Stack Overflow survey.
“The community knows that trust in AI risks causing more problems down the line, not only for themselves but for their coworkers,” Erin Yepis, senior analyst, market research and insights at Stack Overflow, said in an email. “Applying AI for specific, low-risk tasks can be a solution for this. We’re seeing that when AI enhances focused workflows like code reviews or monitoring, it leads to higher sentiment, even if those uses currently apply to a minority of developers.”
Enterprises are bullish on software development boosts from AI. Walmart’s developers have tools that identify accessibility gaps in code and suggest improvements. Food industry giant Mondelēz International has also given its engineers AI-powered tools to shorten development cycles and serve as a learning resource for new hires. Financial services companies have also been early adopters of AI in software development to drive modernization efforts.
