The visceral response to AI interaction is well-documented and physiologically intense. You On AI recounts the experience of watching a functioning application emerge from a conversation — seeing an idea take material form in minutes rather than months. The response is unmistakable: awe, acceleration, a surge of creative energy that changes posture and breathing and sense of what is possible. The collapsing imagination-to-artifact ratio produces sensory impact no previous tool has delivered.
Norman would note that this visceral response, however genuine, is dangerous in a specific way. Positive affect broadens attention, increases tolerance for ambiguity, promotes creative association, and reduces critical scrutiny. These are valuable cognitive states in many contexts. They are precisely the wrong states for the careful evaluation that AI-generated outputs require. The awe suppresses the skepticism. The excitement undermines the caution.
At the behavioral level, AI interaction produces a satisfaction that maps closely onto flow. The iterative cycle of articulation, production, evaluation, refinement has all the structural features flow research identifies. But the flow may be decoupled from skill development in ways traditional flow was not. The person experiences the pleasure of productive engagement without exercising the skills that generated the pleasure. This is what You On AI calls productive addiction, and Norman's framework gives it diagnostic precision: a behavioral-level emotional response decoupled from the growth the response is supposed to signal.
The reflective level — where the person steps outside the experience and asks what she is learning, what kind of practitioner she is becoming, whether the pleasure signals growth or numbness — is where the design challenge is most acute and least addressed. Current AI systems do nothing to support reflective processing. They are designed for engagement, not interruption. They produce output continuously, without pause, without invitation to assess. The feeling crowds out the reflection. The design response the Norman volume proposes includes periodic interaction summaries, transition-point prompts, and dependency tracking — friction at the reflective level that preserves what the behavioral level erodes.
Norman developed the three-level framework in Emotional Design: Why We Love (or Hate) Everyday Things (Basic Books, 2004), drawing on neuroscience research into emotional processing and his own observations that purely functional design accounts had failed to explain why users formed attachments to certain products.
Chapter 5 of the Norman volume extends the framework to AI-assisted work, applying it as a diagnostic lens for a kind of experience Norman's original formulation anticipated but could not have foreseen in its current intensity.
Visceral: immediate, pre-conscious response. The gut reaction to appearance, sound, feel. AI's collapsing imagination-to-artifact ratio produces visceral impact unprecedented in tool history.
Behavioral: satisfaction or frustration from use. The accumulated pleasure of productive engagement. AI engagement resembles flow but may be decoupled from the skill development flow traditionally accompanied.
Reflective: conscious retrospective evaluation. The level of meaning — what the experience says about who the person is becoming. Current AI systems structurally suppress this level.
Design must engage all three. Optimizing for visceral and behavioral satisfaction while ignoring reflective processing produces a design failure of unprecedented sophistication.