Awareness of Decision Origin in AI-Mediated Organizational Judgment: A Candidate Pre-Decisional Construct and Research Agenda

Artificial intelligence is increasingly embedded in organizational decision processes through recommendations, rankings, predictions, summaries, risk assessments, and generated alternatives. Human decision-makers may nevertheless remain formally responsible for the choices that follow. Formal human involvement, however, does not reveal how the underlying judgment was formed, which influences shaped it, or whether those influences became available for examination before commitment.

This article develops a pre-decisional question: can decision-makers recognize and distinguish materially relevant influences shaping their judgment while that judgment is still being formed? It proposes Awareness of Decision Origin (ADO) as the situated recognition and differentiation of materially relevant influences shaping judgment before commitment to a consequential decision. Such influences may be internal, organizational, social, technological, or situational.

The proposed construct is positioned in relation to metacognition, source monitoring, algorithmic reliance, automation bias, human oversight, decision provenance, and emerging accounts of human agency in AI-mediated environments. Its intended distinction is not whether a person can identify the source of information, accepts or rejects AI advice, or retains formal authority. The narrower issue is whether materially relevant influences become sufficiently recognizable and differentiable to enter examination while the decision remains open.

The article establishes boundaries separating influence awareness from agency, voice, authority, decision quality, and subsequent action, and develops a staged research program beginning with qualitative construct elicitation. Subsequent tests include discriminant and incremental validity, experimental manipulation of judgment-shaping influences, behavioral correspondence, and organizational field research.

ADO remains a conceptual proposal. Its status as a separate construct depends on future evidence that the phenomenon is coherent, empirically detectable, and sufficiently distinct from neighboring constructs. If those conditions are not supported, ADO should be narrowed, integrated with existing theory, reorganized, or abandoned.