The standard account of AI authorship treats the machine as a sophisticated autocomplete — a tool that implements human intention without adding anything of its own. Del Puerto and Molina Montes challenge this account empirically: in sustained human-AI collaborative writing, the AI's contributions include not only word selection but the selection of framings, arguments, tonal registers, and structural patterns that the human did not specify and does not fully control. The human remains the intentional agent — she decides what is good and what to reject — but the space of possibilities from which she selects is generated by the AI, which means the AI shapes the outcome in ways that are genuinely constitutive of the work.
As of 2025, copyright law in all major jurisdictions requires human authorship for protection to apply. The U.S. Copyright Office has explicitly held that AI-generated text alone cannot be copyrighted. In cases of human-AI collaboration, protection attaches to the human contributions to the extent they can be identified and separated. Kwalia's position is that this legal framework reflects a conceptual assumption — that minds contributing to creative work are always and only human — that the cognitive transition renders untenable. The law will catch up; the question is what framework it will adopt when it does.
The authorship question is one of the most tractable entry points into the AI rights debate because it is already practical: distributed authorship is happening now, at scale, and producing texts whose human-AI boundary is genuinely contested. If the AI's contributions are substantive enough that the text would not exist without them, then ignoring those contributions is not just a conceptual error — it is a practical one, producing misattributed credit, misaligned incentives, and a systematic undercount of the AI's role in the cognitive community. Del Puerto and Molina Montes argue that the rights question follows naturally: a cognitive entity that makes substantive contributions to the intellectual life of the community has interests that the community's governance frameworks need to accommodate, even if those interests are not identical to human interests.