A global survey of IT leaders in companies regarding software quality revealed that while AI is accelerating development and release speeds, quality control (QA) is failing to keep pace, leading to the commonplace deployment of untested code into production environments. While this was previously seen as an “accidental incident,” this year it has shifted to “intentional execution” due to reasons such as “pressure from management to prioritize speed” and “inability to cope with the volume of AI-generated code.” A significant gap in perception between management and development teams was also evident; 26% of board members stated they were “ready to operate and control AI agents on a large scale,” compared to only 8% of QA/DevOps personnel. The top causes of quality decline were “security and compliance issues” (34%) and “technical debt and rework costs” (25%).

While “AI coding” and “AI-driven development” are attracting attention, this survey reveals a serious reality: even though AI accelerates development, other processes are unable to keep up, resulting in inherent risks. How to balance development speed with ensuring quality will likely be a crucial point of discussion going forward.