Designing ethical AI literacy across secondary to graduate STEM pathways: ensuring cross-disciplinary outcomes and workforce alignment.

Abstract

Rapid adoption of generative AI in STEM courses has created an urgent need for practical, ethical AI literacy that students can carry from high school into college and beyond. In this Perspective, we make the case for a cross-disciplinary approach to AI literacy and present a skills progression table that aligns core outcomes with typical academic and professional levels (high school, early undergraduate, advanced undergraduate, graduate, and workplace). We focus on six core areas: (i) attribution and citation, (ii) bias awareness and mitigation, (iii) privacy and data handling, (iv) safety and environmental care, (v) community impact and fairness, and (vi) compliance and readiness for audit. We frame AI-related classroom practice as a balancing act, on one side, clear benefits (faster drafting, coding support, and broader access), and on the other, real risks (bias, privacy leaks, over-reliance, and academic integrity concerns). To ground our Perspective in STEM broadly, we offer guidance with suggested steps and actions, grading cues, rural/low-bandwidth adaptations, and links to workforce expectations. We also note concrete examples from biology, engineering, and computer science to show both shared core practices and discipline-specific needs. As we look into the future, we highlight the role of professional societies as conveners to ensure alignment of guidance across disciplines and keep materials current. Our goal is to provide instructors with practical materials they can adopt, while building a living framework that evolves with evidence, standards, and stakeholder input.