Carnegie Mellon University

Eberly Center

Teaching Excellence & Educational Innovation

What’s the Eberly Center’s position on adapting teaching to generative AI (GAI) tools? 

GAI can be viewed as a challenge and/or an opportunity for students and instructors.

In some contexts, GAI has the potential to enhance learning outcomes, student equity, or instructor efficacy/efficiency. However, in others, it may not, or actually be detrimental to learning, equity, or academic integrity. For a particular teaching context and learning objective, often, any of the outcomes above are plausible (see also the Eberly Center FAQ on GAI). Ultimately, the impacts of GAI may depend on the underlying teaching and learning principles that are at play given a particular implementation. Regardless, rigorous data directly comparing learning outcomes with and without student use of GAI are lacking.

Consequently, we do not view GAI as a “hammer” and we do not think every teaching and learning challenge is a “nail.” In other words, we are NOT advocating that instructors use GAI just to use it. Instead, we advocate an intentional, student-centered, inclusive, and evidence-based approach to adapting to GAI in teaching (described below). 

Eberly colleagues welcome the opportunity to discuss the pros and cons of different possible strategies for adapting your CMU courses. Email us at

The impact of GAI on higher education is an open “teaching as research” question and the Eberly Center is here to help instructors.

Because we do NOT yet know how its use will affect student learning or equity, our approach is to explore the impacts of genAI through action research (i.e., research to inform practice). Our goal is to help CMU instructors collect, analyze, interpret, and disseminate data that directly measures the impacts of GAI on student learning and attitudes, rather than relying entirely on intuition and/or theory to inform practice. We define data broadly, but it refers to direct measures. And, we always consider and test the alternative hypotheses that implementing GAI may increase, decrease, or not affect student outcomes. Leveraging our expertise in quantitative and qualitative education research methods, our mission is to make it easier for instructors to leverage a data-informed, iterative approach to exploring innovations in teaching and learning.

Contact us at to discuss your ideas for innovating with and collecting data on the impacts of GAI tools in CMU courses. 

Adoption and study of GAI in teaching must navigate ethical concerns. 

We consider potential implementations and studies of GAI with an open mind, but with caution because there are conspicuous ethical concerns related to educational applications, depending on the use case, tool, and implementation (e.g., privacy, accessibility, data security, IP, costs to students, legal implications and more). With any use case or study, we seek to mitigate concerns and avoid harm.

Learning objectives, backward design, and direct measures of student outcomes should guide explorations of the impacts of GAI.

We advocate the following heuristic:

  1. Think intentionally about a learning objective and formulate a research question about it with respect to GAI. How might it affect student learning or equity? 
  2. Then, design how one could collect data to test that question in a CMU course (i.e., learning with vs. without AI use OR which AI use works better for learning). What are possible sources of data directly measuring student learning or attitudes? What comparison groups are possible to infer causality in observed data? 
  3. And finally, ideate and select options for how you might implement a gen AI tool to test your research question.
  4. What are possible implications for diversity, equity, inclusion and belonging?

We acknowledge that conducting the thought experiment above might result in your decision to NOT pursue the research question or implement gen AI in your teaching. Similarly, results of studies might influence whether or not you continue to employ GAI in your teaching. However, these choices would be OK with us because they would be the result of an intentional thought process.

And, if you end up NOT pursuing a research question about GAI or do NOT use GAI at all, the Eberly Center can still support your study of other educational research questions or teaching interventions in CMU courses. We have other fellowships, programs, and 1-on-1 consults that are available to support teaching as research as part of your regular practice as educators!

We are here to help: