When OpenAI’s models arrived, plenty of people decided quickly that the software could write anything worth reading. A short prompt, a few minutes and a finished newsletter or blog post, good enough that a person no longer needed to sit down and write it themselves. I watched that conviction spread through my own field, consulting, and for a while I did not doubt it either.
I kept writing by hand anyway. I took that as a personal shortcoming at first, evidence that I was just slower than everyone else moving on to prompts.
What I have watched for years
I have observed this for a long time now, across consulting firms and among authors who write books about artificial intelligence. Consultants recommend AI-generated content to their clients, blog posts, newsletters, marketing copy, and continue to write their own client proposals and reports by hand. AI book authors, myself included, describe AI-written text as high quality inside a book they wrote themselves, word by word.
None of this shows up anywhere a client could check it. The recommender sells the recommendation, together with the training and the software behind it. What the recommender actually does at their own desk stays invisible. In thirty years of consulting I have developed a rule of thumb: “what someone does for themselves tells you more than what they recommend to you.” That rule applies here as much as anywhere else in the profession.
Two standards, one behind the other
My own view of the pattern is this. Consultants and authors treat AI content as good enough for someone else’s audience, the daily flood of newsletters and blog posts that hardly anyone reads past the first three sentences. The moment a name attaches to a text and the text is meant to matter, the standard rises, and the same person switches to writing it by hand. Everyone in the business already senses this, even if hardly anyone says it out loud.
There is a real exception, and it is worth naming precisely so the rest of the argument stays sharp. Forms, automated notifications, standard confirmations: no reader expects a human voice behind these, so origin does not matter. The contradiction sits somewhere else, in marketing text and client communication that is supposed to build trust while running on a tool the person selling it would not use for their own bylined work. In my eyes that is a lie. The seller sells a great deal of efficiency on top of a quality they do not actually rely on.
Where this leaves the argument
“I’m no smarter than other authors, but for me writing has a dimension of reflection, learning and joy in expression, which is why I can keep it fairly clearly apart.” That is the honest version of my own position. What I miss is an open conversation about that dimension. Every discussion I encounter stays fixed on the efficiency of content production, as if a text had no other function worth naming.
I expect the value of handwritten work to rise as the AI-generated flood grows, the way a hand-built watch or a piece of handmade furniture holds its value once machine production makes the ordinary version cheap and common. What I would like to see more of, in the meantime, is resilience for the work that is still actually written by hand.