Teaching with AI across Disciplines: Reframing How We Learn: The UnBlooms™ Framework Presented by Tina Austin at OpenAI Inaugural Higher Education Summit, October 20th, 2025

Teaching with AI across Disciplines: Reframing How We Learn: The UnBlooms™ Framework Presented at OpenAI Inaugural Higher Education Summit, October 20th, 2025

Abstract

The rapid integration of generative AI into education has outpaced existing pedagogical frameworks, creating an urgent need to reimagine how we design meaningful learning experiences while reconceptualizing academic integrity for an AI-integrated world.

When surveyed, the majority of faculty report redesigning their assignments since AI became mainstream, yet most continue to rely on Bloom's taxonomy, a hierarchical model developed for twentieth-century classrooms that assumes learning progresses linearly from memorization to creativity. AI has collapsed this hierarchy entirely. Students are now able to "create" polished outputs without understanding, and the foundational assumptions of sequential cognitive development no longer hold.

Drawing from a decade of teaching across regenerative medicine, computational biology, ethics studies, digital innovation, and AI literacy, I demonstrate that neither traditional Bloom's taxonomy nor its inversion adequately supports learning design in the age of generative AI. Critical evaluation of AI outputs, including assessment of accuracy, bias, limitations, and appropriate application, must become a foundational learning competency rather than an afterthought. Following my work being featured in OpenAI Academy for custom educational bots and presenting at Oxford University's AI in Education conference in September 2025, I have continued to develop and refine UnBlooms™, a non-hierarchical, recursive, problem-centered framework for AI-era learning design developed through iterative practice-based research with over nine hundred educators across university settings and piloted across my own courses in regenerative medicine, computational biology, and AI digital innovation.

UnBlooms™ operationalizes three essential principles that fundamentally shift how we approach learning in the AI era:

Learning must be recursive rather than linear. UnBlooms™ replaces Bloom's pyramid with a circular model where learners enter cycles of questioning, generating, critiquing, and refining at whatever point the problem demands—not where curriculum prescribes. A history student analyzing the Renaissance, a biology student conducting enzyme assays, and a computer science student debugging Python code each enter at different cognitive points, reflecting authentic disciplinary thinking rather than artificial hierarchies. The challenge is determining which entry point serves the learning goal—and how to scaffold recursive movement through cognitive processes to solve a problem without losing coherence.

Assignments must center student voice and personal experience. AI can generate analysis, synthesis, and evaluation instantly—but it cannot replicate lived experience. The framework provides strategies for grounding assignments in what students uniquely know, creating learning spaces where personal perspective becomes pedagogically essential rather than supplementary. The critical question becomes: how do we design tasks that make student voice structurally necessary rather than optionally valued?

Assessment must shift from evaluating outputs to evaluating critical evaluation itself. Rather than grading what students produce—which AI can now generate with minimal cognitive engagement—educators need methods for assessing students' capacity to interrogate AI's limitations, articulate epistemological gaps between algorithmic pattern matching and disciplinary expertise, and design interventions that address those gaps. This requires new assessment architectures that measure metacognition and adaptive problem-solving rather than polished deliverables.

The framework includes the UnBlooms™ Critical Evaluation Scale, which structures how students engage with AI outputs across progressively sophisticated levels of analytical depth, and a Reflective Decision Tree that helps educators determine when AI integration is pedagogically justified versus when it risks automating the curiosity students need to develop. The talk will explore how these tools shift institutional discourse from "how can we use AI?" or " AI as a cheating tool" to "How can AI help with critical thinking?" an important reframing as institutions rush to adopt technology without adequate pedagogical grounding.

Early evidence from classroom implementations demonstrates measurable shifts in how students interrogate AI outputs, articulate systemic concerns about algorithmic bias, and distinguish between their own reasoning and machine-generated content. Workshop feedback from participating educators indicates the framework's utility in navigating the tension between leveraging AI's potential and preserving pedagogical integrity.

As generative AI fundamentally continues to change intellectual labor, education must evolve from knowledge transmission to transformative practice. This talk provides foundational strategies for implementing recursive, reflection-centered learning design across disciplinary contexts. Customized workshops are available for institutions seeking discipline-specific implementation guidance and practical pedagogical support.