Informational vs. Transformational Learning

In the AI Story

The distinction maps onto the difference between training and development. Training assumes a stable self that needs new capabilities. Development assumes a self that must become different in order to integrate the capabilities meaningfully. A developer learning to use Claude Code can acquire the informational knowledge in hours: how to frame prompts, iterate on outputs, integrate AI-generated code into existing systems. This is training, and it is necessary. But using the tool well — directing it toward purposes that matter, evaluating its outputs with judgment, maintaining one's sense of professional identity while the nature of the profession restructures — requires transformation: the capacity to see coding expertise as one form of contribution rather than as the definition of professional self, to generate one's own standards for quality rather than borrowing them from the community, to hold the tension between what the tool makes easy and what remains hard without collapsing into either celebration or mourning. This is developmental work, and it cannot be completed in hours, weeks, or even months. It unfolds across years, supported by relationships and environments that make the anxiety of identity reconstruction survivable.

AI is an informational amplifier of unprecedented power. It delivers vast amounts of knowledge, executes complex procedures, and synthesizes information across domains faster and more comprehensively than any human expert. This capability creates a dangerous illusion: that the AI transition is fundamentally an informational challenge, addressable through better training. Kegan's framework exposes the illusion. The challenge is transformational. The tools demand not that people know more but that they know differently — that they develop the capacity to hold their professional expertise as object (something they use) rather than subject (something they are). The self-authoring mind can begin this shift. The socialized mind cannot, because it lacks the internal architecture to stand apart from the role and examine it. Providing more information to a socialized mind confronting a transformational demand is like providing a more detailed map to someone who has not yet learned to read maps. The information is useless until the capacity to use it has been developed.

Educational institutions are particularly vulnerable to this conflation. The response to AI in schools and universities has been overwhelmingly informational: policies about permissible use, training in prompt engineering, assignments redesigned to be AI-resistant. What is absent is any systematic attention to the transformational challenge AI presents to students' developing identities. The student who uses AI to complete assignments without developing the capacity to generate her own questions, evaluate competing sources, or exercise judgment about what claims deserve investigation is acquiring information without transformation. She is learning to prompt, not learning to think. The outputs may earn high grades — AI-assisted work can be polished, comprehensive, and correct — but the student's order of consciousness is not advancing. She remains at the socialized level, producing what the authority (now the AI as proxy for the teacher) expects, without developing the self-authorship that genuine education should cultivate.

The distinction also clarifies what mentoring and coaching must become in the AI age. The mentor's role is not to provide information — AI handles that more efficiently than any human. The mentor's role is to support transformation: helping the mentee surface the competing commitments that prevent change, creating space for the grief of identity loss, modeling what it looks like to hold one's own expertise as object rather than subject, and staying present through the confusion of transition. This is slow work, irreducible to productivity metrics, and structurally at odds with the culture of speed that AI enables. Organizations that recognize the difference between informational and transformational learning will invest in mentoring as developmental infrastructure. Those that do not will continue to deploy training programs and wonder why the transformation they expected does not occur — missing the truth that transformation cannot be delivered. It can only be supported, in relationships and environments built for the purpose.

Origin

Kegan developed the informational-transformational distinction through his critique of traditional adult education, which assumed that the goal of learning was the accumulation of knowledge rather than the development of new ways of knowing. The distinction became explicit in his teaching at Harvard, where he observed that students could master course content (informational learning) without undergoing the epistemological shifts (transformational learning) that the content was meant to catalyze. A student could ace an exam on Piaget's stages without developing the capacity to think about her own thinking — which was the deeper developmental goal the course was ostensibly serving. Kegan formalized the distinction in In Over Our Heads, using it to explain why well-designed interventions often failed: they addressed information while the problem was transformation.

The concept gained wider currency through adult learning theory, particularly Jack Mezirow's transformative learning framework and the work of Laurent Daloz on mentorship. But Kegan's contribution was to ground transformation in a precise developmental mechanism — the subject-object shift — rather than leaving it as a vague aspiration. Transformation is not any significant learning. It is the specific process by which structures of meaning-making move from subject to object, and that process has identifiable characteristics: it is gradual, relational, emotionally demanding, and supported by holding environments that balance challenge and support. Organizations and schools that confuse the two kinds of learning systematically underinvest in the conditions transformation requires.

Key Ideas

Informational is additive. Content accumulates within an existing meaning-making framework — the self remains unchanged at the structural level.

Transformational is architectural. The framework itself reorganizes — what was invisible becomes visible, identity becomes capacity, the self grows in complexity.

AI delivers information brilliantly. Large language models retrieve, synthesize, and generate informational content at unprecedented scale and speed.

AI cannot deliver transformation. Developmental growth requires relational processes, emotional processing, and time — none of which AI provides or accelerates.

Conflation produces failure. Organizations treating AI adoption as informational (provide training) while ignoring the transformational dimension (support identity work) predictably fail to achieve meaningful change.