Hands-on AI Pedagogy Showcase: The AI Assessment Scale

It has now been over two years since OpenAI made ChatGPT available to the general public. Thousands of meetings, debates, webinars, workshops, articles, social media posts, and TikTok tutorials later, higher education is still adjusting to the rapidly evolving reality of generative Artificial Intelligence (AI). Whether we like it or not, students are using AI tools at very high rates: according to a 2024 global survey conducted by the Digital Education Council, 98% of higher-ed students report using AI in their academic work, with 54% reporting daily or weekly use. How can we help students engage with AI tools and navigate AI-saturated information ecosystems with ethical integrity and critical awareness? What does generative AI’s ubiquity mean for our own teaching practices? How might we reframe our pedagogical work in light of this new reality? (For an introduction, or a refresher, feel free to browse our previous AI teaching tips, including (1) a general primer on the intersection of AI and pedagogy, (2) considerations for AI-conscious assignment design, and (3) recommendations for composing an AI course policy.)  

Since ChatGPT’s debut in November 2022, many educators have begun developing activities, assignments, courses, and even entire curricula focused on application and/or critique of generative AI. This collective labor has yielded helpful tools, use cases, and conceptual frameworks that can be applied in educational work across disciplines. One of these tools is the AIAS or AI Assessment Scale developed by Leon Furze and colleagues (Perkins, M., Furze, L., Roe, J, & MacVaugh, J. 2024). The function of the AI Assessment Scale is to provide students with specific, granular guidelines concerning generative AI use in their academic work.

The AI Assessment Scale provides a convenient categorization of AI uses in student work, while recognizing that AI might be employed in diverse ways at diverse stages of an academic project. Determining appropriate and/or inappropriate uses of AI tools for each individual assignment requires thoughtful consideration of the assignment’s learning goals as well as a clear understanding of its component sub-tasks. As a result, and perhaps most importantly, the AIAS model provides a starting point for classroom conversations about AI’s utility (or lack thereof) for specific tasks within specific disciplines. The current scale (itself a revision of an earlier, less nuanced, model) can be adjusted to fit the needs of individual courses and assignments—a project that might be undertaken in collaboration with students to encourage metacognitive reflection on how AI is changing the way they work and learn.

Importantly, the AI Assessment Scale was not conceived as a punitive tool for surveilling and “catching” AI offenders. Rather, it offers an accessible visualization of AI uses at various stages of an academic task, with the ultimate goal of helping students understand when (and why) AI tools can harm or enhance their learning.