When AI drafts the text, does a byline still mean the named person wrote it, or only that they steered it?

This explores whether the long-standing deal behind human ghostwriting carries over to AI. In that deal the named author takes credit because they steer the work and the voice is supposed to be theirs. One caveat first: the collection has no studies of human ghostwriting norms themselves. What it does have is a sharp picture of the AI half of the comparison, and that picture shows the ghostwriting bargain quietly breaking in several places.

Start with ownership. Research on what's been called the AI Ghostwriter Effect finds that people's sense of owning AI-drafted text rises with how much control they had over it. Personalizing the model to sound like them does nothing Does user control over AI text shape feelings of ownership?. That matches the human ghostwriting intuition: you earn the byline by directing the work, not by having a writer who imitates you well. The problem is that people mostly don't direct the work. Writers edited AI-generated paragraphs only 23% of the time, and their edits left the text about 96% similar to the original Do writers actually edit AI-generated text before publishing?. A human ghostwriter goes through drafts with the named author. AI text usually goes out close to untouched.

That matters because the AI is not a neutral ghost. A study of nearly 3,000 writers and 11,000 readers found that AI help shifted how readers saw the writer on all 29 traits measured. Writers came across as more confident, more extreme, more agreeable and more privileged Does AI writing assistance change how readers perceive the writer?. In a separate experiment, autocomplete pulled Indian writers toward Western phrasing and cultural references Do AI writing assistants push non-Western writers toward Western styles?. A good human ghostwriter's job is to disappear into the client's voice. AI pulls the client's voice toward its own defaults, so what readers credit to 'the author' is partly the model's personality.

Disclosure is where the comparison gets strange. With human ghostwriting, everyone roughly knows the convention (politicians' memoirs, executives' op-eds), so not naming the ghostwriter misleads no one. AI use has no shared convention yet, so attribution has turned into detection. Rewriting may wipe out the signs of authorship, though whether it also fools AI detectors hasn't been tested yet Do rewrites that hide authorship also fool AI detectors?. AI fiction can still be spotted from story-structure choices that surface edits don't remove Can AI stories be detected without analyzing writing style?. Meanwhile, readers' accusations of AI use often land on comments that show no real AI markers, so the people harmed are human writers who get doubted Do unfounded AI accusations harm human writers instead?. On social media, AI posts collect engagement without building any one speaker's reputation. Credit piles up, but there's no one for it to belong to Does AI content displace human influencers on social media?.

The corpus also points to a different kind of attribution. Instead of asking 'who wrote this?', ask 'where did each claim come from?' A newsroom system that links every number and quote to its source won over professional raters because its output could be checked, not because it was polished Can source traceability make AI writing trustworthy?. That may be the more useful norm to borrow from human practice. Ghostwriting is acceptable because the named author stands behind the claims. AI writing may need to make those claims traceable, since the person whose name is on it often hasn't checked them.