Neither banning AI nor adopting it blindly: Gonzaga's C.A.R.E. framework exposes universities' blind spot

By Zendoric · October 7, 2026.

Jay Yang directs the Institute for Informatics and Applied Technology at Gonzaga University, a Jesuit university in Spokane (Washington), where he leads a cohort of faculty in the "xAI Faculty Fellows" program studying how AI affects human formation, students' professional readiness and the institution's Jesuit mission. Yang has also developed C.A.R.E.: an acronym for Clarity, Authorship, Reciprocity and Effects, four areas he reviews when working with AI, each with questions to consider, a positive practice and the potential risks of using AI on a specific project, as Yang himself explains to Gonzaga News.

Yang traces the origin of C.A.R.E. to Ignatian pedagogy and the Jesuit concept of "cura personalis" (care for the whole person), as well as to the PAL framework —Purpose, Agency, Learning— developed by another Gonzaga professor, Anny Case. It is not a rulebook but a habit: the goal, he says, is to internalize these questions until they become second nature, and for his own children and other users to end up developing something similar.

What matters is not the acronym but the diagnosis behind it. Yang rejects both extremes of the university debate on AI: banning it, which he calls "a misguided approach," and pushing it "without a well-thought-out strategy," which strikes him as equally damaging. His fundamental objection is to the framing that equates AI with academic cheating: according to Yang, it "tells society how to misuse or abuse the technology, instead of promoting the right use for human flourishing." He proposes assessing students on the process of solving a problem and not only on the final answer, and creating interdisciplinary AI councils at every university to gather the good practices that, he says, already exist in scattered form at almost every institution. He also distinguishes between generative AI as an aid to thinking —brainstorming, connecting points of view, with proper citation— and agentic AI that acts autonomously, which he considers "another conversation."

As sector context, we have previously noted the gap between student adoption and teacher readiness in the US: most university students already use generative AI routinely, while barely one in four professors feels prepared to integrate it into their teaching. That gap is the ground on which initiatives such as C.A.R.E. appear: they do not replace an institutional policy, they fill the space while one fails to arrive.

Our reading is that the value of this case lies not in the numbers —there is no study, no adoption figure, no evidence that the framework reduces AI misuse— but in what it reveals about the state of educational governance: while university leaderships and boards go on debating whether to ban or permit, it is individual professors who are designing, on their own, the missing layer of judgement. It is a pattern we already see in other areas of AI: regulation and institutional strategy lag behind everyday practice, and it is advanced users who improvise the guardrails. In the medium term, we expect frameworks like this one —or the PAL that inspired it— to stop being personal initiatives and become assessable curriculum, just as happened with digital literacy a generation ago.

This connects with a thesis we have been defending: the bottleneck of the next decade is not the capability of the models, but the human capacity to use them with judgement —knowing when to delegate and when not to—. That this skill is being taught today as a conscious practice, with explicit questions about authorship and effects, is exactly the kind of groundwork that makes possible the long-term scenario we advocate: an AI that amplifies learning and frees up time for higher-level thinking, instead of hollowing out the "productive struggle" —the cognitive effort needed to truly learn— that Yang himself identifies as the real risk of cutting corners with AI without judgement. The short-term risk is that, without that judgement, the technology erodes precisely the capacity it should be enhancing in the long run.