SSS Framework: A Think-First, AI-Second Model for AI-Assisted Writing

It is remarkable how quickly AI has become woven into the ways we think, write, and create, bringing renewed attention to intellectual autonomy and the long-term consequences of relying too heavily on algorithms. In this paper, I argue that effective human-AI collaboration depends on a deliberate sequence: human cognition must come first, before AI assistance enters the process. Inspired by Marcus Aurelius' sphere of control, empirical cognitive science on desirable difficulties, and the growing literature on automation bias, I propose the SSS framework, a structured, three-phase model in AI-assisted writing comprising the following: (1) Set an AI-Free Zone, (2) Struggle, and (3) Spar with AI. The framework operationalizes a think-first, AI-second model that has been independently supported by recent empirical research demonstrating that think-first AI interaction models preserve and enhance independent cognitive performance beyond task completion. Unlike traditional cognitive offloading models or uncritical human-AI collaboration paradigms, SSS framework introduces productive friction as an intentional design principle, countering both automation bias and epistemic erosion. The framework also addresses the ethical imperatives of AI adoption (e.g., algorithmic bias, transparency deficits, deification of technology) through user-centered, agency-preserving engagement. Pedagogical implications and boundary conditions for applying SSS across varying task complexity and user expertise are discussed.