This study illustrates an exploration of using the Generative AI for a Web survey that aims to examine cognitive health. In particular, the survey plans to implement the cognitive health measures in English, Spanish, Chinese, Korean, and Vietnamese, which requires translation as well as adaptation of the measures. Performance is explored using multiple platforms and multiple prompts. The study finds 1) that the AI tools are less than perfect in translation; 2) that the AI performance varies by language with greater struggle in low-source languages; 3) that the AI tools provide a reasonable starting point for translation on which human generated edits will be applied; and 4) that the current AI tools are not able to offer adaptations.
