Domain-Grounded AI Validation Curriculum for Arts and Sciences

Abstract

This presentation proposes a cross-disciplinary faculty curriculum for the course Domain-Grounded AI Validation for Arts and Sciences, designed to help non-engineering faculty responsibly integrate AI into their teaching. The purpose of the presentation is to engage with the USU community on advanced AI topics. The proposed curriculum, based on an evaluation of AI use in Social Work and Healthcare practice, generalizes AI validation practices beyond Social Work and Healthcare to disciplines such as English, Languages and Translation, Political Science, Public Health, Psychology, and Technical Communication & Rhetoric. The course prepares faculty to move beyond basic AI use toward active participation in the AI lifecycle, including defining domain-specific tasks, creating evaluation datasets, developing rubrics and metrics, testing AI outputs, identifying bias and hallucinations, and designing classroom assignments that teach students how to critically assess AI systems.

Building on introductory courses, this 15-week student and faculty development course curriculum begins with a refresher on AI as a sociotechnical system and introduces the AI lifecycle for non-engineers. Faculty then translate disciplinary expertise into AI validation tasks, develop annotation protocols, define domain-specific metrics, and examine bias, fairness, and missing context. Mid-course modules focus on taxonomies, rubrics, source grounding, hallucination detection, and adversarial testing. The final weeks guide participants through human oversight, contestability, classroom assignment design, course-level AI policies, peer review, and final presentation of a course-ready AI validation module.

The key benefit to the university is the creation of a scalable model for responsible AI education across Arts and Sciences. Rather than relying on generic AI literacy or engineering-led validation alone, the curriculum positions faculty as disciplinary experts who can determine whether AI outputs are accurate, ethical, contextually appropriate, and pedagogically useful within their fields. For faculty, the course provides practical tools for using AI in class while maintaining academic standards, protecting disciplinary judgment, and improving student learning. It also supports the development of shared AI governance practices, classroom-ready validation modules, and discipline-specific benchmarks that can strengthen teaching innovation, interdisciplinary collaboration, and institutional leadership in responsible AI adoption. For students, the course provides hands-on experience in working with AI systems, applying critical thinking and grounded sources of information.

Presenters