Friction-Led Intelligence: A First Principles Framework for Agentic AI and Design Thinking
Anurag Kadel · 2025
As AI systems increasingly shape how people learn, decide, and reason, dominant design paradigms centered on usability and efficiency are no longer sufficient. This paper introduces Friction-Led Intelligence (FLI)-a first-principles design framework that operationalizes friction as a signal for epistemic drift, semantic misalignment, and teleological ambiguity in AI-human interaction. Building on prior work in Seeds of Sovereignty, which framed friction as a regenerative epistemic intervention, FLI translates that philosophy into a recursive design loop involving ontological grounding, epistemic reflection, and teleological clarity. Through a speculative mental health companion agent, we illustrate how FLI enables systems to surface friction not as failure, but as feedback for co-evolution. This framework contributes to a growing movement toward accountable, pluralistic, and philosophically grounded approaches to agentic AI design.