A majority of software developers are integrating AI coding tools into daily workflows despite concerns about the accuracy and reliability of AI-generated code.

Security concerns emerge as a large share of developers use personal accounts for AI tools—35% overall, rising to 52% among ChatGPT users and 63% among Perplexity users—potentially exposing confidential information.

Independent research finds AI-generated code contains more bugs and major issues than human-produced code, roughly 1.7 times higher.

Experts emphasize the need for essential knowledge and practical strategies to leverage AI while bolstering security and verification, echoing the shift in software engineering’s value proposition.

Additional challenges include personal AI accounts for work and the ongoing toil of correcting AI code, complicating organizational code verification practices.

AI usage is strongest in prototyping (88%) and internet production software (83%), with substantial use in customer-facing apps (73%); the most common assistants are GitHub Copilot (75%) and ChatGPT (74%).

AI introduction creates a verification debt, moving value from rapid generation to deployment confidence and exposing a trust gap between AI output and deployment readiness.

Many developers spend significant effort verifying AI code or finding that verification takes more time, with 38% agreeing it requires more time and 61% noting AI output often looks correct but isn’t.

Despite heavy use, 95% of developers verify AI output and 59% report this verification as a moderate or substantial effort, with 38% saying reviewing AI-generated code requires more effort than reviewing human-written code.

The Sonar State of Code Developer Survey shows 72% of developers use AI coding tools daily or multiple times a day, and 42% say AI assists in a substantial portion of their code, with expectations rising to 65% by 2027.

Most developers do not fully trust AI-generated code to be functionally correct, with 96% expressing distrust.

Despite distrust, AI-generated code now accounts for about 42% of developers’ code, up from 6% in 2023, and is expected to rise to around 65% by 2027.