Playing with fire: how AI acts as both arsonist and firefighter in transformative learning
Blog post co-authored by Saskia Eschenbacher and Rachel Fichter
As adult educators, the first theory that pops into our minds when thinking about AI is Transformative Learning. Developed by Jack Mezirow in the 1970s, his ideas are shedding light on transformative dimensions of adult learning. Caused by a disorienting dilemma – a powerful experience that challenges your assumptions about how the world works – transformative learning has the potential to fundamentally and dramatically change your life and what it feels like “being you,” with no way back.
Premise reflection – terrifying, dangerous and liberating
Adult educators support learners in reworking crisis into opportunity through what Mezirow refers to as premise reflection, or the “process of becoming critically aware of how and why our assumptions have come to constrain the way we perceive, understand, and feel about our world” (Mezirow, 1991, p. 167). This fundamental questioning and reordering of our foundational assumptions is an “epiphanic, or apocalyptic, cognitive event,” literally “a shift in the tectonic plates of one’s assumptive clusters” (Brookfield, 2000, p. 139).
Such an event often feels dangerous, terrifying, and liberating all at once, due to the radical questioning of one’s assumptions: “terrifying because it means giving up the familiar banisters and guidelines that we normally accept in orienting our lives; dangerous because, when such questioning is truly radical, it seems to leave us with nothing; liberating because it frees us from illusions and enables us to confront our subjectivity and inwardness without illusions” (Bernstein 2016, p. 121, italics in original). The promised outcome, liberation, seems to make pursuing this form of learning worthwhile.
What are we liberating ourselves from? The dead weight of oppressive assumptions and beliefs that work against us. Transformative learning encompasses “learning how we are caught in our own history and reliving it” (Mezirow, 1978, p. 101), so instead of carrying this weight of our limiting assumptions and beliefs, which we have uncritically internalized from others (family, media, society, etc.), we begin to challenge them. But how do we do this?
Mezirow suggests rational discourse with the Habermasian ideal speech situation. We partly disagree. While discourse is well suited for the public sphere when we debate how we want to live together, it is less relevant for the personal sphere, where we ask ourselves how we want to live our individual lives. In this case, discourse and argument do not help pursue an answer to these very personal questions. Going back to Mezirow’s idea of premise reflection, you might then wonder: How does this work? As Brookfield (2009) notes, “We find it very difficult to stand outside ourselves and see how some of our most deeply held values and beliefs lead us into distorted and constrained ways of being” (p. 133). This is particularly difficult when it comes to sociocultural and psychological assumptions. How can we identify and come to understand more about the assumptions we hold vs. the assumptions that are holding us?
The dual role of AI in transformative learning
That is where AI comes in. Here we use a sociomaterial definition of AI, which assumes that knowledge is “contextualized within particular relationships between people, things and spaces” (Bearman & Ajjawi, 2023, p. 1162) and envisions the “human/non-human action and knowledge as entangled in systemic webs…” (Fenwick, 2010, p. 111).
Within this sociomaterial framing, paradoxically, while AI has the potential to create a disorienting dilemma for us by, for instance, challenging what it means to be human, it is, at the same time, an enabler of premise reflection, the so called “antidote” to a disorienting dilemma. This is because AI is exceptionally good at analysing language patterns (Passmore & Tee, 2023), serving as a “critical mirror” (Brookfield, 2000) to help an individual identify their underlying assumptions as the foundation for challenging them. AI can also be a great brainstorming partner and add depth and breadth to the patterns by identifying possible gaps and suggesting ways to fill them.
But that is not the end of the story. This situated and relational epistemological lens suggests that AI “is determined not solely by the technology itself, but by the contextually bound relationship between the person and the technology” (Bearman & Ajjawi, 2023, p. 1163). Furthermore, learning in “an AI-mediated world should orient towards what things do together, rather than what they are separately” (p. 1164, italics in original). Within the context of transformative learning and premise reflection, AI’s pattern-recognition capabilities can trigger “psychologically explosive” (Brookfield, 1990) emotional responses, eliciting “edge-emotions” such as fear, anxiety, or shame that signal resistance to challenging core beliefs (Mälkki, 2012). This points to a need for careful application of this technological potential–and where the relational definition of AI points to the role of the human educator in helping the individual to engage with these boundary-pushing emotions in a psychologically safe way.
Creating synergies in learning for transformation: Hybrid intelligence
If AI does such a good job identifying assumptions, what, if anything, can the human educator contribute to an individual’s learning process? In other words, what aspects of human intelligence are uniquely human? Informed by cognitive science, the 4E framework–embodied, embedded, extended, enactive–offers a useful perspective: embodied cognition proposes that the body “plays a constitutive role in cognition, literally as a part of a cognitive system;” embedded cognition suggests that the “cognitive capacities of an individual are enhanced when provided with the opportunity to interact with features of a suitably organized physical or social environment;” extended cognition refers to “the environmental and social resources that enhance the cognitive capacities of an agent are in fact constituents of a larger cognitive system;” and enactive cognition refers to the idea that cognition emerges from sensorimotor activity or cognition is generated and specified through operation of sensorimotor processes that crisscross the brain, body, and world” (Shapiro & Spaulding, 2025, secs. 2.1–2.4).
How do we bring these two forms of intelligence together to support transformative learning? AI identifies patterns and assumptions, which the human educator taps different forms of cognition to delve into context-sensitive inquiries, facilitating client reflection and emotional processing. The result is a partnership between the individual, the AI, and the human educator–a form of Hybrid Intelligence, in which a synergistic collaboration emerges that leverages the unique strengths of each member of the partnership to achieve better outcomes (Dellerman, 2019).
One thing we know: AI is here to stay. But by simultaneously setting fires and fighting flames, AI can be both disruptor and facilitator, paving the way for a meaningful human-machine partnership in adult and transformative learning.
References
Bearman, M., & Ajjawi, R. (2023). Learning to work with the black box: Pedagogy for a world with artificial intelligence. British Journal of Educational Technology, 54(5), 1160-1173.
Bernstein, R. J. (2016). Ironic life. Cambridge: Polity Press.
Brookfield, S. D. (1990). Using critical incidents to explore learners’ assumptions. In J. Mezirow (Ed.), Fostering critical reflection in adulthood: A guide to transformative and emancipatory learning (pp. 177–193). San Francisco: Jossey-Bass.
Brookfield, S. D. (2000). Transformative learning as ideology critique. In J. Mezirow (ed.), Learning as transformation. Critical perspectives on a theory in progress (pp. 125 – 48). San Francisco: Jossey-Bass.
Brookfield, S. D. (2009). Engaging critical reflection in corporate America, In J. Mezirow & E. W. Taylor (eds.), Transformative learning in practice. Insights from community, workplace, and higher Education (pp. 125–135). San Francisco: Jossey-Bass.
Dellermann, D., Ebel, P., Söllner, M., & Leimeister, J. M. (2019). Hybrid intelligence. Business & Information Systems Engineering, 61(5), 637-643.
Fenwick, T. (2010). Re-thinking the “thing”. Journal of Workplace Learning, 22(1/2), 104–116. https://doi.org/10.1108/13665621011012898
Mälkki, K. (2012). Rethinking disorienting dilemmas within real-life crises: The role of reflection in negotiating emotionally chaotic experiences. Adult Education Quarterly, 62(3), 207–229.
Mezirow, J. (1978). Perspective transformation. Adult Education Quarterly, 28 (2), 100 – 110.
Mezirow, J. (1991). Transformative dimensions of adult learning. San Francisco: Jossey-Bass.
Passmore, J., & Tee, D. (2023). Coaching psychology: Applying integrative coaching within education. International Journal of Leadership in Public Services. 2(2). 27–33.
Shapiro, L., & Spaulding, S. (2025). Embodied cognition. In E. N. Zalta & U. Nodelman (Eds.), Stanford Encyclopedia of Philosophy (Summer 2025 ed.). Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/sum2025/entries/embodied-cognition/
About the authors
Dr. Saskia Eschenbacher is an EPALE Adult Education Expert. She is Professor of Adult Learning and Counselling at Akkon University of Applied Human Sciences, in Berlin. She earned a Ph.D. (2018) in Education from the University of Augsburg. In 2015 and 2018, she spent terms as a Visiting Researcher at New York University’s Steinhardt School of Education, and in 2019, 2020, 2022 and 2023 at Teachers College, Columbia University. As part of her research, and also as a practicing Systemic Therapist and Consultant, Eschenbacher is interested in how to promote and catalyse processes of personal change and transformation. She is a co-editor of The Palgrave Handbook of Learning for Transformation (2022).
Dr. Rachel Fichter is an Adjunct Assistant Professor in the department of Adult Learning and Leadership at Teachers College, Columbia University and guest faculty at the Université Paris II Panthéon-Assas. A scholar and practitioner, Dr. Fichter has held senior positions in Human Resources at Fortune 100 and 500 companies. She received her MBA from Duke University and Doctor of Education from Columbia University, where she was also certified as a leadership coach. She was awarded a German Government Grant (D.A.A.D) for post-graduate studies in Düsseldorf. Her current research explores the relationships between collective leadership, complexity, and sustainability.
Reading this blog had me…
Reading this blog had me thinking about my 11th graders – where AI sometimes plays the tricky role of a double agent in the classroom. On one hand, students may feel intimidated by how quickly AI generates ideas, as if their own creativity is under threat. On the other hand, AI helps uncover assumptions and explore new approaches. In digital design lessons, I encourage students to experiment with AI to spot patterns, spark inspiration, and challenge their own ideas – never to replace them. This aligns with what the authors describe as AI enabling “premise reflection” (Bearman & Ajjawi, 2023; Brookfield, 2000), triggering that moment of realization: “Ah, I didn’t even know I thought that way!” Practically, this nurtures adaptation and critical thinking – essential aspects of transformative learning. By reflecting on patterns and assumptions, students see how their beliefs shape their thinking. My takeaway: AI may spark the fire, but we hold the matches – shaping ideas and growth.