From Impact Promises to System Design: Embedding Human Constraints into AI Systems

Over the past decade, impact, ESG, and Responsible AI frameworks have promised more equitable outcomes, yet a key gap remains: impact rarely enters system design. This session introduces an approach to move beyond principles into design. Using examples from platform labor, we show how human constraints (e.g., income stability, risk exposure, capacity limits) can be embedded into decision logic. In a joint format, participants co-create solutions using a structured tool (“coupling node diagnostic card”), identify measurable outcomes (e.g., retention, income stability, risk reduction), and explore how cross-sector collaboration can drive systems change. Participants leave with a repeatable method to translate impact into system design.