Survey Surfaces Rising Tide of Production Issues Traced Back to AI Code

A recent survey of business and engineering executives reveals a growing trend of production incidents and defects traced back to AI-generated code in the past year. Conducted by Wakefield Research on behalf of Sauce Labs, an application testing platform provider, the survey found that 80% of respondents have experienced production issues linked to AI-generated code.

Over 10% of production code is now generated using AI, with 28% having over a quarter of their production code generated by AI tools. However, only 38% of respondents receive regular reports on this AI-generated code, and only 27% consistently disclose to customers when AI is used.

Nearly 70% of organizations spend over $1 million annually on AI-assisted application development tools, with 41% spending $5 million or more. 89% of respondents believe the return on investment (ROI) in AI is either significantly or somewhat positive, with 51% expecting fully autonomous testing and software deployment within the next two to three years. While 53% believe deploying AI too quickly poses a greater risk than falling behind competitors, 47% think falling behind is the bigger risk.

The survey also found that 84% of respondents lead organizations that have reduced roles due to AI adoption, with 53% cutting junior/entry-level developers, 42% reducing QA testers, and 34% lessening technical writers. Despite these cuts, 64% report an increase in dedicated quality assurance/testing headcount in the past year. The survey suggests that not enough attention is being paid to testing AI-generated code before deployment, with 66% admitting to compromising on quality or testing standards to meet release deadlines.