Contemporary work systems increasingly distribute analysis, execution, monitoring, and decision activity across employees and intelligent technologies, yet organizational scholarship often classifies these relations with broad labels such as automation, augmentation, or human-in-the-loop design. Those labels obscure who actually performs the focal task, who holds consequential authority, whether contributions are reciprocal, and who oversees whom. This Original Taxonomy Article develops an evidence-informed but explicitly unvalidated classification of human–technology collaboration. The proposed taxonomy distinguishes assistance, substitution, oversight, and coworking arrangements and treats hybrid and transitional forms as configurational or temporal combinations rather than as a higher-order maturity stage. The approach separates task execution from decision authority, technological capability from legitimate responsibility, formal human presence from consequential influence, and collaboration form from organizational outcome. Assistance describes bounded technological input with retained human integration; substitution describes transfer of bounded task execution; oversight captures directional monitoring, validation, evaluation, or control; and coworking describes recurrent reciprocal task interdependence. The taxonomy further proposes that classification should be based on enacted work relations rather than technology labels or claimed implementation strategies. Its implications concern work design, authority allocation, governance, measurement, and longitudinal research on movement between arrangements. The synthesis does not establish causal superiority, universal applicability, validated thresholds, or implementation readiness. Empirical testing should examine discriminant validity, multilevel effects, temporal transitions, contextual moderators, and whether the proposed categories explain organizational phenomena beyond simpler automation–augmentation continua.
