When we use artificial intelligence to assist our work—to gather, synthesise, or even phrase ideas—at what point does augmentation become appropriation? The question is much more complicated than it initially seems, because authorship extends far beyond "did my friend actually write my birthday card?" to issues of IP ownership, originality, confidentiality, and public policy. One of the more recent examples is Washington city officials are using ChatGPT for government work. A case which has not only a legal dimension but raises existential questions about the nature of authorship in the digital age. This is because using AI to draft reports, essays, books, and policies begs questions that cut to the heart of authorship, originality, and what it means to think for oneself in an age where machines can emulate the patterns of elementary human thought.
But let me be clear for those who are in doubt—AI has not introduced plagiarism into the world, though it has greatly enhanced the ease and therefore the temptation. However, what it has done is expose how poorly we have defined authorship as a society. The uncomfortable truth is that few works—academic, corporate, or creative—are entirely the product of one author. Most are collaborative syntheses of others' labour, interpretation, and creation. A collaboration which determines is it authoritative or is it authoritarian.
The philosopher Hannah Arendt warned that authority collapses when we forget its source—when we treat power as self-justifying rather than reasoned. The same might be said of authorship: when we treat originality as possession rather than participation, we erode the very basis of intellectual integrity.
Authorship As Reasoned Elaboration
Carl Friedrich, writing in the mid-20th century, defined authority not as coercion but as reasoned elaboration: the capacity to persuade others through coherent argument grounded in shared values. For Friedrich, authority is a source of power, not a form of it—it derives from our ability to justify and make intelligible our claims to others. This insight offers a useful parallel for the question of AI authorship.
To write authoritatively is not about producing words but in owning the reasoning behind them. When a human author uses AI to draft sentences, the moral question is not "who wrote the words?" but "who stands behind the reasoning?" If AI generates text without discernible human oversight, there is no reasoning to elaborate—only the mimicry of thought without the moral burden of belief.
Friedrich's idea illuminates a central paradox of the AI era: the distinction between saying something and standing behind it. To borrow his terms, an AI output without human elaboration may have the form of argument but not its authority. Authorship, therefore, remains not in the generation of text but in the accountability for its meaning.