← The Journal

Method Note

·8 June 2026·3 min read

Using AI without outsourcing professional judgement

AI can help revise wording quickly. Verifying facts, certainty, obligations and context still requires the person sending the message.

By Casey Bawden , Creator of Neutral Authority Method

There is a reasonable question about whether someone who works on the wording of professional communication should use AI tools to help draft it. Doesn’t that outsource the thing being examined?

The short answer is that AI can be a genuinely useful drafting aid, and using it well is different from handing over judgement. The distinction is in how the output is used, not whether the tool is used at all.

What AI can do

AI tools can generate alternative phrasings quickly, flag habitual constructions a writer might not notice in their own draft, adjust register for a different audience, and produce a first pass at a routine message. These are real, useful capabilities, and there is no basis for claiming they don’t exist or don’t help.

What AI cannot verify

What an AI tool cannot do is confirm that a stated fact is accurate, that a date is genuinely fixed rather than provisional, that a commitment is one the sender can actually keep, or that a claim of responsibility matches what actually happened. Those are matters the tool has no independent access to. Only the person sending the message holds that information.

  • , Is the fact in this draft correct?
  • , Does the certainty in the wording match what I actually know?
  • , Does this accurately describe who is responsible?
  • , Is this appropriate for the sector or relationship it will be read in?

7 Emails, free

Seven ordinary professional emails, with the original, one possible revision and the reasoning behind it.

The PDF arrives by email. You can unsubscribe at any time. See the Privacy Policy.

Two ways of using the tool

One way to use AI is to accept its output with minimal review, correct for obvious typos, then send. In that mode, the tool’s default phrasing becomes the sender’s phrasing, defaults and all.

A different way is to treat the output as a draft to be checked against what the writer knows: does this claim hold up, does this qualifier match my actual certainty, does this sound like something I would stand behind if asked about it directly? In that mode, the tool speeds up drafting without replacing the judgement that has to sit with the sender.

Where responsibility still sits

Sector-specific obligations, legal, clinical, regulated, contractual, do not change because a draft passed through an AI tool. Whoever sends the message remains responsible for its accuracy and its consequences, in the same way they would if they had typed every word themselves.

The practical question is not whether to use AI for drafting. It is whether the review step, checking facts, certainty, responsibility and context, still happens before the message is sent.

For a related look at self-review, see why habitual wording is difficult to notice in your own writing.

7 Emails

7 Emails I Would Write Differently Now