Why do readers and writers disagree on disclosure necessity?

The study's most striking result is an interaction. Writers' intentionality in steering AI generation has "contrasting effects on readers' and writers' perceived necessity of disclosure": low intentionality raises readers' judgment that disclosure is necessary and lowers writers' judgment. The authors say they found this "to our surprise." The abstract summarizes intentionality only as a condition that raises necessity, and the conclusion supplies the split. The excerpt gives the direction in words but no effect sizes, test statistics or cell means.

The excerpt offers no mechanism for the reversal. Its perspective section gives reasons each side might care about disclosure: readers may use it to judge how much to trust a text or whether it is worth reading, while writers may see disclosure as a moral act tied to honesty, or fear that it makes AI co-created work look less valuable or the writer look less competent. Any of these could bear on the reversal, but the excerpt does not test which, if any, does. Intentionality is defined as "how intentionally the writer steered AI generation," so the reversal concerns the writer's steering of AI output, a separate factor from directness, which concerns how much AI text ends up in the piece.

The co-writing ownership study, Does ownership framing change how much writers rely on AI?, also ties a writer's stance toward AI output to how they act: owners lean on AI suggestions. Its outcome is reliance, while this vignette measures a disclosure judgment, so the two sit side by side without testing each other. The authorship dissociation note, Do users truly own the AI-generated content they produce?, offers a possible reason the writer side discounts AI's role: writers can declare authorship at a reflective level while feeling little cognitive ownership. This excerpt measures neither authorship claims nor felt ownership, so that link stays a hypothesis. The first note, Do readers and writers differ on AI disclosure necessity?, reports intentionality as one of the conditions that raise necessity overall; this interaction qualifies that reading, since the same signal moves the two perspectives in opposite directions.

The excerpt does not show whether the reversal holds outside vignettes, where participants imagine being readers or writers. The limitations section flags social desirability bias for writer-perspective participants, which bears directly on a result that depends on the writer side. The excerpt also skips the intentionality levels and procedural hypotheses. The implication, at the strength the evidence allows: a signal that makes readers want more disclosure does not simply carry over to writers, so guidance built on one audience's reaction to low steering would not transfer to the other. That is a design question the study raises and does not answer.