Leapwork CEO warns AI‑generated code could amplify software quality debt

Ken Ziegler, chief executive of Leapwork, cautions that the rise of AI‑assisted code generation may dramatically increase software quality and security debt. While AI tools can produce syntactically clean code at a faster rate, they often embed false assumptions that go undetected until production failures occur. Ziegler explains that the problem lies in a lack of validation: "The asymmetry held because creation is visible and validation isn’t," and stresses that without systematic testing, AI‑generated code can reinforce its own errors.

A Leapwork survey cited in the article shows that 88 % of development professionals view AI as a priority, yet only 12.6 % currently employ it for key testing activities, and 59 % of testing remains manual. Ziegler advocates a shift toward “continuous validation,” integrating AI‑driven testing with robust data layers and executive oversight to prevent the rapid expansion of hidden defects.

The CEO argues that enterprises must start with trusted data, prioritize where AI adds value, and avoid applying AI to flawed processes, which merely amplifies existing problems. He suggests practical steps such as hackathons, consolidated data repositories, and balanced human‑AI collaboration to achieve reliable production software.