The Human Cause

Why It Matters

We are accustomed to machines performing physical labor that humans once did. That transition, whatever its costs, left the domain of the mind intact — the writing of novels, the composition of music, the generation of scientific hypotheses, the design of new molecules. These activities were still, distinctively, human. Artificial intelligence is now entering that domain, and entering it fast.

This is not merely a legal problem. It is a civilizational one. Literature, music, and journalism are among the most important vehicles by which cultures understand and reshape themselves. If the works that fill those vehicles are produced by machines, then machines will be steering — however inadvertently — the arc of culture and, ultimately, of political life. Call it self-driving culture. The metaphor is meant to unsettle.

The US Constitution is unusual among constitutional documents in stating not just the power to legislate on copyright and patents, but the purpose of that legislation: to "Promote the Progress of Science and Useful Arts." The question this scholarship takes seriously is whether that clause means human progress. The answer given here is yes — and it has implications for how IP law should respond to the rise of generative AI.

"Change is a difference between two points A and B on a timeline, while progress is an improvement at point B. Change happens; progress not necessarily."

Human self-fulfillment and realization through art and science is progress. It is a position shared, in various forms, by Plato's Laws, Aristotle's Nicomachean Ethics, and Spinoza's concept of conatus. The law should aim to foster those goals — not merely reward the production of outputs, regardless of their origin.

The IP Stakes

Copyright

Copyright law, as it has developed in both common law and civil law traditions, has always presumed a human author. The Berne Convention's protections flow from acts of human creative expression. Most national laws, under comparative examination, arrive at the same conclusion: only works made by humans are eligible for copyright protection.

The challenge posed by AI is not hypothetical. Machines already produce outputs that are often indistinguishable from human-authored works — poems, news reports, musical compositions, fiction. If those outputs receive copyright protection, the economic consequences follow inescapably: since machines do not demand royalties or hold reversion rights, market forces will favor machine-produced content wherever it is commercially feasible. Human authors will lose the financial flows that enable them to develop their craft. The "muscles" of human creativity, used less and less, will atrophy.

This research distinguishes two questions that tend to be conflated: whether an AI output resembles a copyrightable work, and whether it actually qualifies for protection. On the argument developed here, the resemblance is precisely the problem — it is a new kind of Turing test, rewarding the owner of a machine that can pass itself off as human. That illusion is not a sound normative foundation for IP protection.

Patents

In patent law, the question of inventorship is similarly foundational. Inventorship is not merely a formality — it determines the right to obtain a patent and who benefits from the invention's commercial exploitation. Courts have consistently held that inventors must be human. Yet as AI systems become capable of generating genuinely novel and non-obvious technical contributions — in pharmaceutical research, materials science, and beyond — the question of whether a machine can be a named inventor is being litigated directly.

The argument here is that the answer should remain no — not because machines cannot produce valuable technical outputs, but because IP rights exist to reward and incentivize human inventive activity. Granting patent rights on AI-generated inventions would redirect the rewards of the patent system to the owners and operators of AI systems, while providing no additional incentive for the machines themselves, which have no interests to incentivize.

A Causation Test

How, then, should the law distinguish outputs with a human cause from those without? The scholarship developed here proposes drawing on the law of causation — familiar from tort law — as an analytical tool.

In tort law, causation asks whether a human act was a cause-in-fact of a given outcome. The question for IP law is analogous: was there a human act — a genuine creative or inventive contribution — that was a meaningful cause of the work or invention for which protection is sought? The test is not whether a human was anywhere in the process (any prompt to an AI system involves a human), but whether that human contribution rises to the level of authorship or inventorship as those concepts have historically been understood.

This framework allows a spectrum. At one end: a purely machine-generated output, with no meaningful human contribution beyond deploying the tool. No protection. At the other: a work in which AI is used as a tool, in the way a writer uses a word processor or a composer uses notation software, with the creative choices made by a human author. Full protection. In between: a range of cases requiring judgment, but with the causation framework providing principled guidance.

The framework is technology-neutral — it does not require the law to distinguish between one AI system and another — and it is consistent with the historical foundations of both copyright and patent doctrine.

What Progress Really Means

The deepest argument is philosophical. "Progress" in the constitutional sense — and in the sense that has animated IP law across two centuries and dozens of legal systems — is not the same as the production of more outputs, faster, at lower cost. Progress, on the account offered here, is human progress: the growth of human understanding, the flourishing of human expression, the expansion of the range of human self-realization.

Delegating creative and inventive activity to machines does not advance that kind of progress. It may, in fact, reverse it. As we rely more heavily on machines to perform tasks that are singularly important to us as a species — generating stories that help us understand our world, composing music that expresses what cannot be said directly, formulating the hypotheses that extend the reach of science — we shrink the space available for individual and collective human self-realization.

The more we outsource those activities, the more, as a species, we may lose on two fronts: the diminishment of human expression itself, and the erosion of the financial and social conditions under which human creators and inventors develop their craft. Both losses matter. Both are, on this account, things the law should be concerned to prevent.

The human cause, in the end, is a good cause in itself — not a sentimental attachment to the past, but a reasoned commitment to a particular vision of what the future, at its best, could look like.