Research

1. The CDAI Index (Curiosity–Doubt–Agency Index)

• Research output: A diagnostic evaluation framework for benchmarking curiosity, productive doubt, and user agency across AI interfaces.

• What it measures: How specific interaction patterns influence exploration behavior, willingness to question and verify, confidence calibration, and perceived agency.

• Rationale: Many AI interfaces create an "illusion of mastery"—high user confidence paired with shallow understanding. CDAI is designed to make curiosity-preserving design measurable and auditable.

2. Reflective-AI Design Playbook (Pattern Library)

• Research output: A tested library of low-friction interface patterns that move AI systems from "answer engines" toward inquiry-driven, reflective partners.

• What it includes: Patterns such as Socratic scaffolding, verbalized uncertainty, and structured alternatives—designed to counter single-answer bias while preserving usability.

• Rationale: Interface design choices can either deepen learning and reflection or accelerate dependency on fluent, frictionless answers. This playbook translates research evidence into practical, deployable product guidance.

3. Public Synthesis & Narrative Translation

• Research output: High-impact communication artifacts, including interactive essays and short videos.

• What it does: Translates complex technical findings on cognitive authority into accessible, narrative forms for practitioners, policymakers, and the public.

• Rationale: To ensure research findings reach beyond academic silos and influence the broader cultural norms and "civic imagination" of the AI era.

4. Inquiry Labs for Educators

• Research output: A deployable pedagogy package designed for middle school, high school, and undergraduate contexts.

• What it includes: Teachable, repeatable protocols and micro-lessons that fit real-world curricular constraints.

• Rationale: To prevent AI-enabled "shallow mastery" and equip students with the cognitive habits required to think with AI rather than simply through it.

5. Deference Conditions Atlas

• Research output: A field-intelligence map linking real-world usage contexts to specific cognitive failure modes.

• What it identifies: The psychological and environmental triggers—such as time pressure, high stakes, or perceived fluency—that lead users to "premature closure".

• Rationale: Understanding precisely when and why humans defer to AI is a prerequisite for designing calibrated interventions that restore human agency.

6. Evidence Maps & Evaluation Synthesis

• Research output: An audit and "Gap Note" connecting findings across HCI, cognitive science, education, and AI ethics.

• What it does: Synthesizes the existing state of research to identify high-leverage unknowns in how AI systems influence human judgment and independent thinking.

• Rationale: To anchor all project interventions and prototypes in rigorous, field-grounded evidence rather than speculative opinion.