The ability of AI systems to provide different results for the same question makes them unsuitable for verifying data in regression tests. This inconsistency undermines the reliability of AI’s assessments, making it difficult to trust their conclusions. Regression testing, on the other hand, relies on consistent and repeatable results to ensure that new features don’t break existing functionality. As AI’s output can vary, it’s not a reliable tool for this critical task.
Therefore, I believe that in the future, we will have regression tests with fixed implemented logic and fixed input and expected data.
