Beyond Output: What Counts as Thinking in the Age of AI?

Brett Christie, Ph.D.
VP, Educational Innovations

Reflections on a webinar with Tina Austin, the UnBlooms™ framework, and the growing reconsideration of Bloom’s Taxonomy in higher education.

Yesterday I had the opportunity to host a timely and thought-provoking webinar with Tina Austin, Beyond Output: Teaching Discernment and Human Reasoning in the Age of AI using UnBlooms™, which drew 278 live participants. At the center of the session was a question many educators are now confronting with greater urgency: if generative AI can produce polished, fluent, and often convincing work in seconds, what actually counts as evidence of student thinking?

Why This Question Matters Now

That question reaches far beyond academic integrity. It is a pedagogical question. It asks us to reconsider what we value in learning, what we design for, and what we choose to assess. For many faculty and faculty developers, AI is not simply disrupting assignments. It is pressing on long-held assumptions about rigor, cognition, and the visible signs of learning.

Why Bloom’s Is Being Reconsidered

That helps explain why Bloom’s Taxonomy has re-entered the conversation in such a visible way. For decades, Bloom’s has served as a familiar reference point for thinking about outcomes, learning activities, and assessment design. What is striking now is not that Bloom’s is being abandoned, but that it is being reconsidered.

Oregon State University’s Bloom’s Taxonomy Revisited, for example, explicitly uses Bloom’s as a “touchstone” for reconsidering course outcomes and student learning in the age of generative AI. The resource is intended to help faculty and instructional designers reflect on activities, assessments, and, where possible, outcomes, while emphasizing both distinctive human skills and ways GenAI can supplement learning processes.

Other reconsiderations are emerging as well. Michelle Kassorla’s Understanding Inverted Bloom’s argues that “create” should not be confused with mere production, especially when AI can generate output so easily. In her framing, the meaningful intellectual work continues after creation through evaluation, analysis, application, understanding, and memory.

Tina Austin’s UnBlooms™ model enters this conversation with a more substantial reframing. Austin describes UnBlooms™ as a non-hierarchical, recursive, problem-centered framework developed in response to the reality that traditional linear models no longer reflect how students think and create with AI. She emphasizes that the framework replaces hierarchy with recursion and repositions reflection as an organizing principle rather than an afterthought.

As Tina Austin argues in her Substack article The Evolution of Bloom’s Taxonomy, And Where It Was Always Headed (May 11, 2026), AI has exposed the limits of treating finished products as sufficient evidence of learning; the more important questions concern what students noticed, questioned, accepted, rejected, revised, and when AI may have hindered learning.

What Our Audience Told Us

That framing made the session especially compelling to me. Rather than treating AI as simply another tool to accommodate, or as a threat to police, Tina pushed us toward a more foundational question: How do we design for observable reasoning? If polished products are no longer reliable enough on their own, then educators need stronger ways to surface judgment, revision, decision-making, critique, and reflection.

Our audience polls suggested that this concern is already widespread. In the first poll “Where has generative AI most changed how you think about assessment?” (multi-response option)

• 79% of respondents said AI has most changed how they think about assignment design and prompts. 
• 71% said it has changed how they think about what counts as evidence of student thinking. 
• 53% pointed to when and how AI should be allowed 
• 43% pointed to how much process students should show. 
• Only 2% said AI had not significantly changed their thinking.

The response pattern suggests that the pressure educators are feeling is not isolated to one part of teaching and assessment. It is spreading across the full design chain, from prompts and activities to process, evidence, and acceptable use. The center of gravity, though, is especially important: faculty are not only asking whether students are using AI. They are asking what kinds of assignments still give them meaningful insight into student reasoning.

The second poll showed that this is not just a conceptual shift. It is already shaping practice. “ Which statement best describes your current approach? (select one) 

• 36% of respondents said they are substantially redesigning assignments to make reasoning more visible
• 31% said they are making small adjustments to assignments or instructions. 
• 26% said they are still figuring out what changes are needed 
• Only 4% said they are mostly keeping existing assignments and expectations in place
• Just 3% said they are not yet addressing AI in a deliberate way.

These responses suggest that educators are not only rethinking assessment in principle, but actively revising practice in response to AI’s growing influence.

From Policing to Pedagogy

The webinar chat added even more texture. Participants described shifts toward writing conferences, video reflections, think-aloud audio files, before-and-after walkthroughs of AI-assisted revisions, and assignment structures that make cognitive moves more visible rather than evaluating only a finished product. The chat also reflected a clear values orientation. Participants emphasized “thinking and processing growth, not perfect output,” noted that “the answers are not the point,” and insisted that “doing your own thinking has intrinsic value.”

At one point in the session, I placed a simple phrase into the chat: from policing to pedagogy. It clearly resonated. The responses that followed pushed the conversation away from detection and toward design. One participant replied with a call for the community to “Please focus on process versus AI detection tools,” while others pointed to the limitations of automated detection and the need to keep attention on reasoning, judgment, and learning design instead.

That, for me, was one of the most important signals from the session. The more useful question is not how to prove whether AI touched a piece of work. The more useful question, going back to Austin and UnBlooms™ is what kinds of learning experiences require students to interpret, justify, reflect, decide, and revise in ways that remain meaningfully their own.

The chat also surfaced harder unresolved questions that deserve continued attention. Participants raised issues related to critical AI literacy, collaborative learning, deliberation, and student decision-making. One participant warned that when decision-making is surrendered to AI, part of the learner’s growth is surrendered with it. Another raised a particularly important challenge: if we are shifting toward process, exploration, and mastery-oriented learning, how do we help students see that this work is just as real, rigorous, and worthy of assessment as a traditional final paper or polished end product?

What Higher Education May Need Next

I appreciated that the session did not pretend these tensions are fully resolved. If anything, that honesty made the conversation more useful. The rethinking of Bloom’s Taxonomy in the age of AI, by Austin and others, is pressing the issue further by centering recursive reasoning, reflection, and problem-centered learning design.

One of my clearest takeaways from the session is that AI is not just forcing us to revise assignments. It is forcing us to sharpen our educational commitments. If the product no longer tells us enough, then we have to become more intentional about the kinds of human thinking we want to make visible. That means designing for discernment, not just completion. It means creating space for decision points, critique, revision, and justified judgment. And it means recognizing that in an age of easy output, the deeper work of higher education may lie less in what students can produce and more in how deliberately they think along the way.