Reputation-filtered: authorship is part of the product

What this profile actually means

Names filter attention because no one can evaluate everything from scratch. Your audience cannot read every essay, test every tool, or vet every recommendation. So they use a shortcut: they trust the name. They opened this because it was yours, expecting the judgment, taste, or standard that the name has come to mean.

That shortcut is the product. The artifact is not being judged purely on its own merits; it is being judged as a sample of you. Which means assisted production is acceptable only on one condition: the named judgment is still legible. If the name opened the door, the named person has to be visibly in the room.

The specific risk: forwarding instead of filtering

The failure mode here is subtle. The work can be accurate, useful, and polished, and still betray the reader, because they were not buying accuracy. They were buying your filter. If what they get is competent output you merely forwarded, the name made a promise the work did not keep.

This is different from the relationship breach, where the harm is false intimacy. Here the harm is a diluted signature. The reader trusted that your taste pre-selected this, that your standard was applied, that the stranger observation or sharp stance was yours. Generic assisted output, passed along under a name that means something, quietly spends down the reputation that made the name worth following.

How to keep AI subordinate to the signature

You do not have to avoid AI. You have to keep it underneath a visible act of taste. Make the named judgment the load-bearing part:

• Take a sharper stance than generic output would. AI hedges toward the consensus. Your name is worth following because it does not. State the thing you actually believe, including the part that costs you.
• Show the selection. What you included and, more tellingly, what you cut. Curation is taste made visible. An audience that follows a name is largely paying for the editing.
• Add the stranger observation. The detail, preference, or connection only you would make. AI produces the expected; your filter exists to surface the unexpected.
• Own the accountability. Put your name on the call, the prediction, the recommendation. Being willing to be wrong in public is part of what the reputation is built on.

A concrete example

A respected reviewer publishes a roundup of tools. If it reads like a competent summary anyone could have assembled, the reputation leaks, even if every fact is right. The reader wanted the reviewer's verdict: which one they actually use, which one disappointed them, which trade-off they would make. The fix is not to hide the AI that helped draft the descriptions. It is to make the taste unmistakable: "I dropped two popular options from this list on purpose, and here is the one I switched to myself." The assistance recedes because the signature is loud.

When this is the wrong profile

If the audience is not actually filtering through your name and just wants the result to work, you may be outcome-first. If the draft is generic before the question of authorship even arises, close the specificity gap first. If the work addresses people personally rather than trading on reputation, see Relationship breach risk. And if reputation is in play but a clean note covers the exposure, see Trust-sensitive.

The principle to keep

A name is an attention shortcut backed by a track record. Every artifact published under it either deposits or withdraws. AI can help you produce more, but it cannot supply the taste that made the name worth a click. Keep the assistance under the signature, and make the signature impossible to miss.