Zway.ai

Answer

Is ghostwriting for founders ethical?

Yes, when the account is what it claims to be. A founder's account carrying a founder's own approved views, drafted by someone else, is the same arrangement as every executive speechwriter and every ghostwritten CEO letter. An invented persona presented as a real person is not, and that is the actual line.

Last reviewed 27 August 2026

The principle

An account has to be what it claims to be. That is the whole test, and almost every hard case resolves against it cleanly.

A founder's account claims two things: that a specific real person is accountable for what appears there, and that the views expressed are that person's views. It does not claim that the person typed every character, any more than a keynote claims the speaker wrote the speech, or an annual shareholder letter claims the CEO drafted it alone, or a book with one name on the cover claims no editor touched it.

Those arrangements are so old and so normal that we have institutions built on them. Nobody thinks a president is a fraud for having a speechwriter. The relevant question was never authorship of the sentences. It was whether the person standing behind them means them.

The strongest objection, stated properly

The best argument against founder ghostwriting is not that drafting is inherently dishonest. It is that the analogy to speechwriting smuggles in the ethics of one medium and applies them to a different one.

A speech is visibly institutional. Everyone in the room understands they are watching a prepared performance by an organisation. A founder's social account promises something else entirely. It trades on intimacy. The implicit offer is unmediated access to a particular mind at a particular moment: what this person noticed this morning, what they actually think about a competitor, the thing they got wrong last year. The whole reason founder-led content outperforms brand content is that readers believe they are getting the person rather than the company. Ghostwriting that is a bait and switch, the argument goes, even when the views are genuine, because the format's real product is presence rather than opinion.

That is a serious argument and it identifies something real. But it proves less than it sets out to.

Why it does not hold, and what it does establish

The objection conflates two promises the format makes. One is about whose views these are and who is accountable for them. The other is about specific factual claims contained in the post.

The first survives drafting intact. If a founder reads a post, agrees with it, publishes it under their name and can defend every line of it in a room full of customers, the reader has exactly what they were promised: this person's position, with this person's name on it and this person's reputation at risk. Nothing material changes when they learn a colleague wrote the first pass.

The second does not survive, and this is what the objection actually establishes. "I was thinking about this on the walk in" is a claim about the world. If there was no walk, that is a lie. It is a small lie and it is still a lie, and it is completely avoidable, because the post works just as well without it.

So the objection does not make ghostwriting unethical. It correctly identifies that the intimate format raises the cost of inventing specifics, which is a rule about content rather than about who typed it.

What actually crosses the line

Invented people. An account presented as a real employee who does not exist. This is the clearest violation and it is not a grey area.

Fabricated first-person specifics. Anecdotes that did not happen, experience nobody has, opinions constructed for engagement rather than held. The account is then not what it claims to be even though the person is real.

Undisclosed impersonation in live conversation. Someone answering replies as the founder, in real time, without the founder's knowledge or approval. The reader in that thread believes they are talking to a specific person. They are not.

Publishing views the founder has not read. Volume without approval. This is the most common failure in practice, and it is usually how the other three start.

The arrangementAcceptableWhy
A writer drafts, the founder reads and approvesYesThe views and the accountability are the founder's
AI drafts, the founder reads and approvesYesSame arrangement, different drafting method
A post invents a personal anecdoteNoIt makes a false claim about the world
Someone replies live as the founder, unapprovedNoThe reader believes they are talking to a person
An account presents a colleague who does not existNoThe account is not what it claims to be
Posts go out that the founder has not readNoNobody is actually standing behind them

The practical version

Approval before publication, every time, with the founder actually reading it. No invented specifics: if the story did not happen, cut the story rather than improving it. Replies either come from the founder or come from a clearly identified company account. And a standing test that the founder could defend any post on the account, unprompted, in front of a customer.

Run it that way and the arrangement is ordinary. Skip the approval step and you have not built a voice, you have built a liability with someone else's name on it.

Related questions

Do I have to disclose that someone drafts my posts?
There is no general obligation to, for the same reason a CEO does not footnote the annual letter with the name of the comms lead who drafted it. The obligation attaches to the substance: the views must be yours, you must have read and approved the post, and any specific factual claim in it must be true of you.
Is it different if AI writes the draft rather than a person?
Ethically, no. Both are drafting. Practically, yes, because AI drafts arrive faster than most people review them carefully, and unreviewed publishing is where the real failures happen. Some platforms also have their own rules: Hacker News prohibits generated text outright, and individual communities increasingly restrict it.
What about replies and comments?
This is the part most people get wrong. A drafted post is a considered statement. A reply in a live thread is a conversation, and the reader reasonably believes they are talking to the person whose face is on the account. Someone else answering in your voice without your knowledge is a different act from drafting, and a worse one.
Where is the line in one sentence?
If a reader learned exactly how the post was made, would they feel misled about something that actually mattered to them? Learning it was drafted almost never clears that bar. Learning the person does not exist, or the story did not happen, always does.

Zway runs growth marketing execution for founders. AI plans it, a trained operator ships it.

See how it works