Where do writers locate authenticity in AI co-writing?

Hwang et al. find that writers who co-write with AI conceive of authenticity largely as the process and experience of making the work, not only as the text that results. The abstract says these conceptions "focus more on the process and experience of constructing creators' authentic selves." The evidence is qualitative: semi-structured interviews with 19 professional writers, who co-wrote with a personalized and a non-personalized AI writing tool and were interviewed after each session, supplemented by an online survey of 30 avid readers. The discussion summarizes that writers ground authenticity in "the source of content, internal experiences and identities that ground their work, and the actual writing outcomes."

The reasoning departs from the framework the literature offers. Prior work uses three themes, Category, Source and Value. The authors report that participants' reflections "did resonate with some of the definitions" but "placed further emphasis on viewing authenticity through their internal experiences in addition to through their explicit expressions in writing." The Source theme, read as a "who" question, is where AI registers as a threat. Writers who held that view "saw AI writing assistance as a direct threat to authenticity," because with AI in the process, "writers are no longer the sole source of content generation." The excerpt adds that several internal states are key constructs of authenticity that "have been less covered by prior literature."

Against the nearest notes, this places authenticity inside the writer's experience rather than in the product. Do users truly own the AI-generated content they produce? describes a gap between felt and declared authorship and treats it as structural rather than dishonest. Hwang et al.'s writers make the felt process central to what authenticity is, which suggests a gap of that kind would matter to them as a loss, though the excerpt does not test this. Does ownership framing change how much writers rely on AI? measures reliance by who owns the final product; a product-level account would miss the process-level conception here. The contrast with Does AI writing assistance change how readers perceive the writer? is one of object: that study measures how readers perceive the text, while Hwang et al. ask what the writer experiences as authentic.

The excerpt does not establish how widely this conception holds. The 19 interviews are a qualitative sample, the excerpt gives no counts per theme (the summary table is not included), and the reader survey appears only in the abstract. The abstract also says writers wanted personalization that targets "writers' growth and go beyond the phase of text production," which the excerpt reports without testing against outcomes. The implication is narrower than a general claim: the process-and-experience conception is a well-supported description of how these professional writers talked about authenticity, and a hypothesis about writers in general.