The Function Got Redefined Underneath the Work: Labor-Layer Extraction in AI Development and the Architecture That Could Bind It

SI-WP-010: The Function Got Redefined Underneath the Work The Function Got Redefined Underneath the Work: Labor-Layer Extraction in AI Development and the Architecture That Could Bind It (SI-WP-010) documents the labor-layer manifestation of the trajectory diagnosed at field scale by SI-WP-009. It identifies a function-redefinition mechanism operating at the production layer of AI rather than at the deployment layer where existing governance work has focused, and it assesses the comparative near-term feasibility of five candidate routes for binding architecture against the pattern. Three cases ground the labor-layer pattern: The Oracle case: An engineer instructed to record her workflows before being laid off and replaced by recordings, alongside large-scale layoffs at a company reporting record growth. The supply-chain case: Data workers training AI systems reporting median annual earnings of $22,620, with 22 percent experiencing homelessness, contracted through vendor networks that insulate deploying companies from labor accountability. The Hangzhou case: A Chinese court ruling that AI adoption is a voluntary business strategy rather than an objective circumstance justifying termination, in a case where a quality-assurance worker had refused a 40 percent pay cut after his work was deemed automatable. The three cases share a single mechanism: function-redefinition under competitive pressure, operating beneath the level at which deployment-layer governance instruments can reach. The paper engages the strongest alternative explanations and industry counteraccounts, including productivity-transition, macro-null, AI-washing, competitive-necessity, and contractor-flexibility framings, and argues that these constrain rather than defeat the layer-specific claim. It then assesses five candidate routes for binding architecture: compute and cloud gatekeepers, procurement coalitions, treaty blocs, insurance and liability architecture, and licensing-as-architecture. Licensing-as-architecture has the highest near-term feasibility because it is authority-adjacent to the function-definition layer itself, not merely leverage-adjacent to the actors in the field. Methodological positioning: This paper is pre-empirical, theoretical, and methodological. The labor-layer pattern documentation is well-supported by primary sources. The prescription is structurally specified and architecturally tractable, but has not been deployed at scope. The near-term-feasibility claim for licensing-as-architecture is held to the regulated-profession subset only; the unregulated supply-chain labor layer remains exposed absent additional architecture. The paper does not predict that any candidate route will be taken; it specifies what would have to be true for binding architecture to be deployed. Document ID: SI-WP-010 Version: 2.7.3 Author: Thomas W. Gantz Affiliation: Synthience Institute License: CC-BY 4.0 v2.7.3 update some citation URLsFor published work and Institute information: synthience.org