When learners open AI before forming a single thought of their own, the cognitive work of understanding gets skipped entirely. What follows feels like learning. It is not. Research across seven preregistered experiments shows that learners who receive LLM-generated summaries develop shallower knowledge than those who engage in active construction first. The mechanism is not the AI tool. It is the sequence.

This session introduces the Map Before Machine™ Card, the deployable tool of the Somagraphic Learning™ Framework, a human-first, AI-supported approach grounded in embodied cognition, cognitive load theory, and human-AI interaction research. The framework's three-stage cycle, Attempt → Map → Refine, positions visual reasoning as a structured interface between human thinking and AI output.

Prototype interviews with STEM graduate learners show a clear delta. Both participants produced more specific, gap-based AI prompts after completing the card compared to their unstructured baselines. Both independently identified the Map stage as the highest point of cognitive friction and the strongest pull toward opening AI. Neither was prompted to say this. Both described the logic of the intervention before being told what it was.

The visual orientation stage is pre-verbal by design. Shapes, spatial arrangements, and motion cues do not require English language proficiency to produce or interpret.

For learners navigating complex concepts in a second or third language, AI-first instruction risks building understanding entirely on someone else's conceptual language. This is a structural problem that plays out daily across MENA classrooms. Somagraphic Learning™ addresses it before the AI interaction begins.

The card requires no special tools. It fits inside any existing course workflow in under five minutes. It is printable or LMS-ready for Canvas, Moodle, and equivalent platforms.

Attendees leave with a deployable tool, a clear evidence-based rationale, and a facilitation model ready to test in their next session.

Toprani, D. (2026, April 22). Somagraphic Learning™ Framework: A human-first, AI-supported visual cognitive approach. OSF Preprints. https://doi.org/10.35542/osf.io/fnk7z_v2