MAP-TM: A Proposed Methodological Framework for Multi-AI Probabilistic Trajectory Mapping

This working paper proposes Multi-AI Probabilistic Trajectory Mapping (MAP-TM), a methodological framework for exploring possible trajectories in complex ecosystems through human-directed orchestration of multiple large language models (LLMs). It extends the Unknown quadrant of the Johari Window into a conceptual Johari Frontier, treating the unknown as a space for exploring variables, relations, and alternative trajectories. MAP-TM separates linguistic exploration by LLMs from the computation of event probabilities, which is performed only when justifiable parameters are available. This paper is a methodological proposal. It reports no empirical results; validity, inter-rater reliability, and the effectiveness of multi-LLM orchestration have not yet been tested. A pilot design with five comparison conditions, five falsifiable hypotheses, and a reproducibility protocol is specified for future evaluation.