Strategic AI-human collaboration for biology assessment design: a practical workflow

Biology instructors increasingly use generative artificial intelligence (AI) tools to create assessment rubrics but lack systematic workflows for ensuring scientific accuracy in AI-generated criteria. This article presents a five-phase workflow for strategic AI-human collaboration tested across undergraduate biology courses, including lab reports, genetics problem sets, and discussion assessments. The workflow balances AI efficiency in generating initial criteria with essential human expertise in scientific reasoning, graph interpretation, and discipline-specific evaluation. Key phases include (i) human-led learning outcome definition, (ii) AI-assisted criteria generation, (iii) human validation with biology-specific refinements, (iv) collaborative testing with sample work, and (v) human-led implementation and iteration. The workflow provides decision rules, validation strategies, and field-tested examples for maintaining scientific rigor while leveraging AI efficiency.