AI commoditizes standardized cognitive work and inflates premiums for verifiable human presence; policymakers should treat human-provenance verification as critical labour infrastructure, not a luxury label.

Main Finding

AI-saturated markets will compress the scarcity premium on standardized middle-tier knowledge work by making “good-enough” synthetic substitutes cheap and scalable. In response, a narrower set of human labor will gain value because verified human presence itself becomes scarce and signaling—what the authors call the human-provenance premium. Because access to that premium depends on credible verification, human-provenance verification should be treated as labor infrastructure (designed to preserve worker bargaining power, portability, privacy, and auditability) rather than as a consumer authenticity label.

Key Points

• Mechanism: marginal-cost compression

  • Foundation/agentic models convert many cognitive and creative tasks into infrastructure with high fixed costs and near-zero marginal cost, lowering the buyers’ baseline willingness to pay for human time on standardized tasks.
  • This weakens the scarcity rents that supported middle-tier knowledge occupations, not necessarily collapsing employment but hollowing out the rent-bearing quality of those roles.

• Conditional barbell hypothesis (stylized value reallocation)

  • Value concentrates at two asymmetric poles: (1) large rents to owners of AI/compute/platform infrastructure (synthetic-scale control), and (2) high unit-premium but capacity-constrained roles whose value depends on verified human presence.
  • The middle (commoditized cognitive/creative outputs) shrinks in rent-bearing value.

• Performative humanity and three types of human-presence value

  • Performative humanity: labor whose value depends on visibly human inputs rather than solely on output characteristics.
  • Type I — Relational presence labor: value derives from intersubjective, chosen human attention (e.g., therapy, high-trust advising). Verification question: was meaningful human participation present? Main risk: emotional extraction and surveillance.
  • Type II — Aesthetic provenance labor: value derives from human origin/trace (e.g., handmade arts). Verification question: is the human-origin claim credible? Main risk: synthetic mimicry and fraud.
  • Type III — Accountability labor: value derives from a human bearing legal/professional liability (e.g., physicians, lawyers). Verification question: who bears legal/professional risk? Main risk: liability laundering.

• Constitutive human presence standard

  • Human labor retains premium when human judgment, attention, authorship, accountability, or relational participation is constitutive (not incidental) to what the buyer is purchasing. This avoids a brittle human-vs-machine binary and helps evaluate hybrid human-AI work.

• Human-provenance verification is distributive and infrastructural

  • Verification determines access to premiums; therefore provenance systems should be designed to protect worker bargaining power (portable credentials, privacy-preserving, auditable).
  • If left as consumer-facing authenticity labels, verification may exacerbate power asymmetries (platform capture, credential hoarding).

• Empirical and institutional signals

  • Micro evidence (labelling/auction experiments, productivity exposure studies) establishes preconditions: willingness to pay shifts for human-labeled outputs; AI raises productivity and compresses within-occupation skill premiums.
  • Macro empirical confirmation of a Veblen/upward-sloping demand effect for human provenance remains a forward-looking prediction.

Data & Methods

• Paper type: position/theoretical paper built from literature synthesis and conceptual argumentation (no primary dataset).
• Methods:
  • Literature review across labor economics (task-based polarization, skill-biased change), information economics (near-zero marginal cost of information goods), platform capitalism, sociology of valuation, credence-good theory, and empirical studies of AI exposure and provenance effects.
  • Conceptual taxonomy (performative humanity → three types) and conceptual modeling (barbell-shaped value-capture curve).
  • Cites empirical/experimental studies to support elements of the mechanism:
    • Productivity/exposure studies: e.g., Noy & Zhang (ChatGPT reduced task time), Brynjolfsson et al. (AI assistance raised customer-support productivity).
    • Provenance/valuation experiments: studies showing higher valuations for “handmade” or “human-authored” labels and devaluation when believed to involve AI (e.g., Bellaiche et al., Mandel & Imas).
    • Platform and labor literature documenting concentration of returns to infrastructure owners and distributional patterns in care work.
• Limitations acknowledged by authors:
  • The barbell/Veblen prediction is conditional and forward-looking; direct economy-wide evidence of upward-sloping demand for human-made goods under AI saturation is not yet established.
  • Historical precedents (industrialization of handicrafts) caution against automatic premiumization; outcomes will depend on market depth, certification costs, buyer coordination, and institutional design.

Implications for AI Economics

• Market structure and distribution

  • Expect reorganization of value rather than uniform job collapse: infrastructure owners capture scalable rents; a narrow subset of human labor gains high per-unit premiums but cannot absorb all displaced workers.
  • Distributional risk: premiums may not accrue to current incumbents of relational or care roles (frequently women, migrants), producing inequitable outcomes unless governance intervenes.

• Measurement and research priorities

  • Empirical tests needed: (a) willingness-to-pay and price-elasticity for verified human provenance across sectors; (b) market depth—how many livelihoods can the premium sustain; (c) who captures the premium (workers, platforms, certifiers); (d) effects of different verification technologies on labor bargaining power.
  • Track interaction between regulatory regimes (liability rules, professional licensing) and accountability labor premiums.

• Policy and governance recommendations

  • Treat provenance verification as labor infrastructure:
    • Design verification systems to be portable across platforms and employers, privacy-preserving, and auditable.
    • Ensure verification enhances worker bargaining power (not just consumer signaling).
    • Regulate verification markets to prevent capture by platform owners or certification firms that could siphon premiums away from workers.
  • For accountability labor: clarify legal liability rules so responsibility isn’t laundered through AI; preserve human accountability where social value depends on it.
  • For relational and aesthetic markets: set standards for what constitutes constitutive human presence and how verification supports ethical labor practices (avoid exploitative emotional extraction or surveillance).

• Risks and trade-offs

  • Verification infrastructure can simultaneously protect worker premiums and enable surveillance/commodification of personhood—design choices matter.
  • Synthetic mimicry and fraud may erode provenance value unless verification is robust and broadly trusted.
  • Narrow premium sectors may become positional/Veblen goods, intensifying inequality and limiting the number of workers who can benefit.

• Broader economic theory implications

  • The paper links task-based automation, information-economics cost structures, and positional/credence-goods theory to argue for a new mode of value-formation (performative humanity)—suggesting refinement of models that predict labor polarization under AI.
  • Introduces constitutive human presence as an operationalizable criterion for evaluating whether human labor is complementary to or substituted by AI.

Recommended next steps for researchers and policymakers: - Empirically estimate premiums for verified human presence in representative markets. - Experiment with different verification architectures (portable credentials, zero-knowledge proofs, decentralized attestations) and measure impacts on worker outcomes. - Model long-run labor-market equilibrium under varying assumptions about verification cost, buyer coordination, and regulatory constraints on liability.