Do private AI flags actually change how people write?

The announcement pairs the public button with a flag aimed at posters. LinkedIn "will begin privately flagging in users' dashboards when people believe their content is coming off as inauthentic due to heavy use of AI." The company's stated belief is that this "will help posters improve their writing, in the case that they're simply using AI technology to refine their own work", as distinct from "full-on 'slop' content." The stated split is between a writer refining their own draft, who can be helped, and a poster producing slop, which the button and classifiers handle instead. The excerpt names no measure for the first group beyond the flag itself.

The trigger is described in terms of belief. Flags go out "when people believe" a post is inauthentic, and the excerpt does not say whether "people" means readers, the platform's detectors or both, nor how heavy AI use is measured. The flag is private, which the excerpt reports without giving a reason. It sits beside a second change: LinkedIn is pulling its "enhance your post" feature, which "had used AI to help you write", and replacing it with a feature that proofreads "instead of changing your voice." Read together, the two changes separate correcting a draft from rewriting it, and the stated design goal is to keep the writer's voice. The excerpt does not say why the enhance feature was pulled or how far the proofreading goes.

Two nearest notes bear on this. Do writers want to see each other's AI prompts in shared editors? finds writers wanting AI use visible to their collaborators. The LinkedIn flag runs the other way: the disclosure reaches only the poster, and readers see nothing. Does AI writing assistance change how readers perceive the writer? reports that AI assistance distorted every measured dimension in a study of N=2,939 writers and N=11,091 readers. That gives a voice-protecting choice like proofreading a research basis, but the excerpt does not test the proofreading tool, so it neither confirms nor refutes the switch.

The open question is whether a private flag changes how people write, or only how they report or hide their AI use. The excerpt reports the company's belief and no outcome data: no before-and-after rates for flagged posters, no flag volumes, and no evidence that writers revise, post less or change their disclosure after a flag. Until such data exists, the flag is a design bet that the announcement states as a belief. A reader citing LinkedIn's approach should attribute the claim to the company, not treat it as evidence that writing improves. The record of posting behavior before and after flags would be what settles it, and the excerpt does not contain it.