According to a recent Stack Overflow survey, trust in AI-generated outputs among developers has significantly declined from 2024 to 2025. While AI adoption rose from 76% to 84%, the percentage of developers who distrust AI results surged from 31% to 46%. Only 3.1% of developers express high trust in AI outputs, with trust levels skewed towards novices (6.1%) over seasoned professionals (2.5%). Notably, 78.5% use AI infrequently—monthly at best—across all experience levels. Despite skepticism, many see AI tools as valuable for learning, with 44% using them to learn coding, up from 37% the previous year. However, 75% rely on humans for verification when unsure about AI responses, especially concerning ethical and security concerns. As Stack Overflow CEO Prashanth Chandrasekar notes, while AI can facilitate workflows, its inherent risks necessitate a blend of human oversight for effective application.
Source link
