AI's plan-recognition subfield frames itself as abductive, in a tradition the field traces to Charniak (1985), independent of Wikipedia

abductive-reasoningplan-recognitionach-cdpeircecharniakai-historyargument-from-silence

Robert P. Goldman, Christopher W. Geib, and Christopher A. Miller's "A New Model of Plan Recognition" (Proceedings of the 15th Conference on Uncertainty in Artificial Intelligence, UAI-1999, pp. 245–254) states its own contribution plainly: "We present a new abductive, probabilistic theory of plan recognition." The paper situates this within a named disciplinary lineage: "To the best of our knowledge, Charniak was the first to argue that plan recognition was best understood as a specific form of the general problem of abduction, or reasoning to the best explanation (Charniak & McDermott 1985)." The priority is assigned to Eugene Charniak personally, with the 1985 textbook he wrote with Drew McDermott — Introduction to Artificial Intelligence (Addison Wesley) — cited as where the argument appears. Plan recognition, as the same paper defines it, is the process of inferring an agent's goals from observations of its actions. The paper then reports the formalization: "Charniak and Goldman (C&G) (1993) argued that, since plan recognition involves abduction, it could best be done as Bayesian inference" — a reference to their Artificial Intelligence 64(1):53–79 paper "A Bayesian model of plan recognition."

This classification predates, and is independent of, Wikipedia's "Analysis of competing hypotheses" article — the only other source the vault had found linking plan recognition to abduction (see claim-ach-cd-abduction-characterization-traces-to-wikipedia-editorial-gloss) — by roughly two decades, and comes from within the field's own technical literature rather than an encyclopedia editor's gloss. "Abduction" here means the same thing claim-inference-logical-types records for Peirce: inference to the most plausible explanation of the evidence. The field is not unanimous on this framing: Kautz & Allen's 1986 "Generalized Plan Recognition", the other major foundational theory, explicitly frames the same problem as deduction over circumscription and never uses the word "abductive."