What Makes Creation Human? Authorship, Reasons, and Meaningful Human Control in Generative AI

TEXT START: Generative artificial intelligence (GenAI) significantly expands creators' productive capacity, but this does not necessarily entail a corresponding increase in creative agency or authorship.

The Dissection

The paper is constructing a legitimacy test for human authorship after generative AI destroys the old proxy of manual labor. It rejects four crude standards—physical intervention, automation level, initial ideation, and final selection—and replaces them with Dynamic-Reflexive Tracking: authorship exists when evolving human reasons are reflectively absorbed, causally effective, and capable of rejecting or redirecting the system’s trajectory.

Its real function is to preserve a human category inside machine-mediated production. It makes authorship depend on the quality and continuity of judgment rather than on how many operations the person performs.

The Core Fallacy

The paper’s framework may be coherent as a theory of authorship. It is irrelevant as a defense of mass economic participation.

Under the Discontinuity Thesis, a human can retain genuine reasons, exercise reflective judgment, and author a work while becoming economically unnecessary. DRT protects attribution and agency at the level of the creative act; it does not restore the wage, bargaining power, ownership, or consumption circuit that AI severs.

The deeper error is treating meaningful control as a moat. A person may redirect a model, but the model and its owner still control the production infrastructure, distribution, scale, and capture of value. Human authorship can survive as a credential while human labor becomes disposable. DRT is a control certificate, not a claim on the capital that makes the certificate valuable.

Hidden Assumptions

• Human reasons remain sufficiently opaque, scarce, and non-automatable to confer durable advantage.
• The ability to reject or redirect an AI output translates into control over the resulting economic value.
• Institutions and platforms will recognize DRT-authored work rather than collapse it into cheap machine supply.
• Reflective uptake and trajectory efficacy can be observed and verified without becoming performative compliance theater.
• A creator has meaningful access to models, data, distribution, and capital rather than merely operating as a replaceable interface layer.
• Human-generated reasons cannot be anticipated, induced, optimized, or simulated by increasingly capable systems.
• Recognition of authorship produces material leverage rather than merely symbolic status.
• A traceable human process remains valuable when millions of systems can generate acceptable substitutes at near-zero marginal cost.

These assumptions are precisely where the framework collides with P1, P2, and P3. Once cognitive systems can model human preferences and institutions cannot preserve human-only economic domains at scale, reflective control becomes a niche property, not a general defense.

Social Function

Primary classification: partial truth and transition management. Secondary classifications: prestige signaling and ideological anesthetic.

The paper correctly identifies that human authorship cannot be measured by keystrokes or by the mere existence of an initial idea. But socially, it offers institutions a survivable story: humans remain authors if they can demonstrate a sufficiently meaningful relationship to the machine’s output. That may help allocate credit, establish accountability, and preserve cultural meaning for a minority of creators.

It does not answer who owns the productive systems, who captures the surplus, or what happens when authored work is no longer economically necessary. It refines the etiquette of displacement while leaving displacement intact.

The Verdict

This is a sharp taxonomy of authorship, not a rebuttal to obsolescence. DRT may preserve human meaning and authorship for Sovereigns, elite creators, and indispensable Servitors. It cannot preserve the post-WWII labor order.

The paper diagnoses who may still claim the corpse’s name. It does not make the corpse alive.