Human-Scale AI

In The You On AI Field Guide

The framework generates specific prescriptions at multiple institutional scales. At the organizational scale: the AI Practice framework (structured pauses, sequenced workflows, protected mentoring time) as an attempt at synergic-satisfier design. At the educational scale: curricula that teach understanding alongside production, reflection alongside output. At the labor-institutional scale: retraining at the speed of displacement, economic safety nets adequate to a transition measured in months, regulatory frameworks that give workers meaningful voice. At the cultural scale: protection of leisure as a fundamental need, not as the absence of productivity.

The framework also generates a specific evaluation discipline. For any AI deployment, the question is not 'does it work?' but 'what kind of satisfier is it, in this specific ecology, for this specific community of users?' The same tool can function as synergic for one population and inhibiting for another. The same tool can shift between categories as the surrounding ecology changes. The classification is empirical, ongoing, and requires the kind of sustained assessment that dashboards optimized for quarterly reporting cannot provide.

Human-scale AI is ultimately a political and institutional project, not merely a technical or individual one. It requires the construction of institutions that do not yet exist — independent assessment bodies, community-governed deployment frameworks, regulatory structures that measure what markets cannot price. The question is not whether to build AI. The question is whether to build the institutional infrastructure that would make AI serve a life rather than consume one.

Origin

The concept synthesizes Max-Neef's Human Scale Development framework (1991) with the specific diagnostic challenges of the AI transition (2024–2026), integrating the Berkeley AI Practice research, the broader post-growth tradition, and the emerging literature on AI governance and cognitive sustainability.

Key Ideas

Nine-meter instrument panel. AI evaluated against the full spectrum of human needs, not productivity alone.

Synergic-satisfier goal. Practices and institutions that meet multiple needs simultaneously.

Context-dependent classification. The same tool can be synergic in one ecology and inhibiting in another.

Not anti-AI. Accepts the genuine creation-satisfaction and insists on serving the other eight needs too.

Institutional project. Requires infrastructure that does not yet exist and will not emerge from market dynamics alone.