Developers more productive with AI but only 18 percent fully trust it
New research finds that 85 percent of developers report higher productivity with AI, yet only 18 percent say that they fully trust it.
The study from Techreviewer was conducted among senior developers, CTOs, and tech executives from 19 countries. It shows that 64 percent of developers use AI tools every day, with only two percent never using them. ChatGPT leads in popularity at 84 percent usage, followed by Claude (64 percent), Copilot (56 percent), and Cursor (53 percent).
Top use cases for AI include debugging (62 percent), code generation (58 percent), documentation (53 percent), and testing (53 percent), which are delivering the biggest time savings -- 85 percent report higher productivity, and 64 percent note improved code quality.
Overall satisfaction stands at 78 percent, signaling AI's permanence as a ‘pair programmer’ that boosts efficiency without replacing human judgment.
Only 18 percent say they are fully confident in AI accuracy. 62 percent always verify generated code manually, and 64 percent spend as much or more time on reviews as on original coding.
There are ethical dilemmas too, 80 percent have faced at least one issue, primarily related to intellectual property (46 percent), data privacy (46 percent), and bias (43 percent). Disclosure to clients is inconsistent too, with 20 percent never sharing information about AI usage.
Some complex tasks are still human-led -- architecture planning (18 percent) and API integration (24 percent) remain mostly manual, underscoring AI's limits in context-heavy areas.
You can see more, along with practical recommendations for using AI in development effectively, on the Techreviewer blog.
Image credit: BiancoBlue/Dreamstime.com