Artificial intelligence (AI) promises faster, cheaper credentialing item development, but can it meet the rigor required for valid, fair, defensible tests? In this interactive session, participants will blindly review multiple-choice items, without knowing whether they were created by AI or by experienced item writers. Using criteria from Haladyna, Downing, and Rodriguez (2002), participants will evaluate stem clarity, distractor plausibility, and cognitive level and vote on who or what wrote it. This activity will spark discussion about where AI works well, where it falls short, and why human expertise remains essential in the item-writing process. Also, facilitators will connect these observations to key NCCA guidance on use of AI in certification programs. Participants leave with strategies for responsible AI integration that safeguards validity and fairness, and a clearer understanding of why human expertise remains vital to sound item development.