Skip to content

Relationship breach risk: show the human judgment

The artifact speaks to people personally, as if from you. When the words sound intimate but the human judgment behind them is invisible, the method stops feeling like a tool and starts feeling like a substituted relationship.

What this profile actually means

Personal address creates a relationship. A founder note, an apology, a message to subscribers, a thank-you to a patron, a direct word to a community: each one borrows the warmth of a person speaking to a person. The more the artifact leans on that closeness, the more the audience expects an actual human presence behind it.

The breach risk is specific. It is the moment the audience thought they were hearing from you and later realizes they were mostly hearing a tool-shaped proxy. The content can be polished and still cause harm, because the harm is not about quality. It is about a felt promise of intimacy that turned out to be hollow.

Why this is worse than ordinary disclosure

In most cases, a small collaboration note is enough to keep trust intact. Not here. A footnote saying "AI assisted" does not repair a message that asked someone to feel personally addressed. The intimacy was the thing being delivered. If the human was not really present, a disclaimer just confirms the absence.

This is also where the "transparency dilemma" bites hardest. Telling people that AI wrote the warm note they received does not reassure them. It retroactively cheapens the moment. The repair cannot be a label appended afterward. It has to be human presence built into the message itself.

What "show the human judgment" looks like

You do not need to expose every assisted step. The audience does not care that AI helped you phrase a sentence. They need to see the judgment that survived the tool: the part only a person who actually showed up could provide.

  • Name what you specifically know. In a customer apology, the detail of what broke and what you saw in the logs proves you looked, rather than generating sympathy.
  • Name what you chose. "I decided to refund everyone affected, including the ones who did not ask." A decision carries a person; a sentiment does not.
  • Name what you will do, and own it. Who is responsible, what changes, who answers if it happens again.
  • Name the cost to you. What you regret, what this is going to take, what you are sitting with. Real stakes cannot be faked into existence by fluent prose.

The honest alternative: lower the intimacy

Sometimes you cannot, or do not want to, put real human presence into the message. That is allowed. The honest move is then to stop pretending. Drop the personal address and make the artifact operational instead.

A status update, a clear policy, a factual notice: these do not borrow intimacy, so they do not breach it. "Service was down from 2 to 4pm; affected accounts have been credited; here is what changed" asks nothing of the relationship and breaks no promise. The failure mode is only the false closeness. Remove the closeness and the risk goes with it. The rule of thumb: if the human sender cannot actually stand inside the message, do not write it as though they are.

When this is the wrong profile

If the work is not really personal and the audience just wants the result, you may be outcome-first. If the relationship is real but lower-stakes, a precise note may be enough. See Trust-sensitive. If the draft is also generic, fix the specificity before the intimacy. See Specificity gap.

The principle to keep

Intimacy is a promise. When an artifact addresses someone as a person, it promises a person is there. You can keep that promise by making your judgment visible, or you can release it by lowering the intimacy. What you cannot do is collect the warmth of personal address while leaving the human out of the room.