CONCEPT
Developmental Design
The principle that human-AI collaboration should be structured to develop the human's capability over time, not just accelerate current output—tools making users <em>better</em>, not merely faster or dependent.
Developmental design is Winograd's third and deepest principle for AI collaboration: the interaction should leave the human more capable over time, not less. A tool supporting understanding makes users better at work; a tool replacing understanding makes users dependent—faster at producing results but less capable of evaluating them, less equipped to work without the tool, less able to exercise independent judgment that gives the tool's output value. The language interface presents this developmental question with unusual sharpness: a developer generating code without understanding it has produced output but not deposited understanding. The output may be correct; the developer may lack capacity to evaluate correctness because understanding that would enable evaluation was bypassed in production. Over hundreds of iterations, pragmatic competence rises while foundational comprehension erodes—a pattern Winograd's framework demands designers address through deliberate preservation of formative friction.
In The You On AI Field Guide
The principle connects to Jerome Bruner's concept of scaffolding—temporary support enabling learners to accomplish what exceeds independent capability, designed to be gradually withdrawn as capability
Keep reading with YOU ON AI
Unlock the full book, 10,000+ field-guide entries, and a 1000+ thinker library. If you have a book code, register now — it takes a minute.