The common complaint about AI writing is that it all sounds the same. That is a fair observation, and it is a usage problem more than a tool problem.
Voice disappears when you ask for a finished piece and publish it. It survives when you use AI for the parts that do not carry voice.
Structure is one of those parts. Outlines, section order, what to cover and what to leave out. Research organisation is another. So is the mechanical work of reformatting a long piece for a different platform. None of that is where your personality lives.
Where voice lives is in openings, examples, opinions and the specific details only you would know. Those should be written by you, every time. A useful discipline is to rewrite the first paragraph and the closing yourself, regardless of how good the draft looks.
A second technique is feeding your own earlier writing as a style sample and asking for output closer to it. This works better than asking for a tone in the abstract.
A third is removing the tells. Overused connectors, padded sentences, and grand claims with no specifics are the usual giveaways.
be10x runs AI workshops for working professionals that focus on practical output quality, which is worth exploring if writing is part of your role.
Use AI to write faster. Do not use it to sound like everyone else.
Top comments (0)