Numerous experiments are underway on uses of artificial intelligence in the design and validation of psychological scales, tests and measurements. New developments such as large language models offer novel tools for psychometric tasks such as item generation, validity assessment, and adaptive testing. Researchers have applied these technologies to routine psychometric tasks such as drafting candidate items as well as more nuanced activities such as making content validation judgments. While artificial intelligence holds promise for streamlining instrument development, several pitfalls exist with the potential of adversely impacting important processes such as employee selection and educational testing. This brief review cites recent literature to describe AI-supported content generation and analytics and includes consideration of data quality and ethical issues. The concluding discussion offers three suggestions for improving the responsible use of AI for psychometrics.
Jeffrey M. Stanton (Thu,) studied this question.