AI, Labor Time, and Scarcity: A Prospective Measurement Framework for Competing Marxian Interpretations of Advanced AI

Advanced AI systems can reduce the human time required for some cognitive tasks while increasing output, reorganizing skills, and concentrating control over new infrastructures. These developments revive old questions in Marxian political economy: whether AI should be treated as machinery owned by capital, whether a sufficiently autonomous non-human system could ever become a labor subject, and whether advanced automation instead sharpens a deeper contradiction between material productive capacity and labor time as a social measure. The relevant literatures already contain substantial versions of each argument. This paper therefore does not claim that Marx anticipated large language models, that present AI creates Marxian value, or that the labor-time problem is new. Its contribution is narrower and methodological. It converts three overlapping Marxian interpretations into a dated prospective framework of observable implications while separating theoretical constructs from measurable proxies. Seven hypotheses are mapped to prior support, counterevidence, competing mechanisms, and potential measurements. A secondary scarcity-migration ledger tracks whether monetary gains and bargaining power move toward compute, energy, data, intellectual property, distribution, regulation, capital, or physical infrastructure as cognitive production becomes cheaper. Twelve observation questions are frozen as a 2026 baseline for later updates. The framework treats price, accounting profit, rents, labor shares, labor hours, and ownership concentration as informative but non-identical observables; none directly measures Marxian value or surplus value. The objective is not to prove or disprove Marxism with a single future outcome, but to reduce hindsight reinterpretation by stating in advance what patterns would strengthen, weaken, or fail to discriminate among rival readings.