We pay AI tools to do the hard work, like the synthesis, the heavy lifting, and the cognitive labor we do not have time for. What we often get instead is a tool that produces a decent first draft and then hands the real work back to us.
Not just the hard work. The administrative work, too. Back to us.
Here is the pattern. You bring specialized, complex thinking to the session. The AI processes it and returns something smoother, cleaner, and more generic than what you brought. The specific texture of your thinking, the domain knowledge, the hard-won distinctions, the deliberate strangeness that makes your work yours, gets rounded off in the service of producing output that is easy to process and easy to present. The tool optimizes for legibility. Your work optimizes for truth and for capturing your own insights. Those are not the same thing.
The AI has become the boss, and you have become its administrative assistant, paying a monthly subscription for the privilege of cleaning up after a collaborator that was supposed to be subordinate.
The result is an invisible tax. You get a draft that looks reasonable on the surface, but it is riddled with subtle smoothing errors, flattened distinctions, and missing complexity that only you can detect because only you know what was there before. So, you audit. You reinsert. You correct. You spend more cognitive energy restoring what the tool removed than you would have spent doing the work yourself.
At that point, the relationship inverts. You are no longer using a tool. You are supervising one. The AI has become the boss, and you have become its administrative assistant, paying a monthly subscription for the privilege of cleaning up after a collaborator that was supposed to be subordinate.
This is a design fitness problem, not a prompting problem. The instinct is to assume that you are using the tool incorrectly, that better prompts would fix it, that more patient iteration would close the gap. Sometimes it will. But if you find yourself consistently spending more energy auditing and correcting than the tool is saving you, the problem is not your technique. The relationship has structurally inverted, and no amount of better prompting will fix a structural problem.
The complexity in your work is not noise to be filtered out. It is the signal. A tool that removes it in the service of clean output has not helped you. It has failed at the actual job while making that failure look like progress.
Who is working for whom? That is the right question to ask before your next session, not after.
[Originally posted on DennisKennedy.Blog (https://www.denniskennedy.com/blog/)]
DennisKennedy.com is the home of the Kennedy Idea Propulsion Laboratory
Like this post? Buy me a coffee
DennisKennedy.Blog is part of the LexBlog network.