Does that sound disrespectful? Good. Because it is high time we stopped anthropomorphising – humanising – language models.
Thanking an LLM (ChatGPT, Claude, Gemini or Grok) has no real effect on the quality of its answers. Do not believe me? Ask them yourself. They are not conscious entities. They have no emotions, they do not register appreciation, and even if they did have favourite users, you would not be one of them. They are machines. Computing systems. Mathematics and statistics. Vast, admirable, capable of remarkable results – but still just a computer program.
It is we humans who project humanity even where there is none. So we kneel before algorithms, we thank them, we plead with them, we write „pretty please“ and „thanks in advance“. We believe that politeness will earn us something extra. It will not. Not from machines.
The effect is that a crucial line gets blurred – the line between a person and a tool. And once that line is smudged, we risk no longer asking the right questions: Who trained this model? On what data? What biases does it carry? Whose interests does it serve?
Politeness is a fine human quality. But applied mechanically to tools it becomes a linguistic placebo. It achieves nothing, yet it makes us feel better about ourselves. And that is precisely the trap.
Forget Stanisław Lem, The Matrix and The Terminator. Wonderful books and films – but so are Harry Potter and The Hobbit. And surely nobody in their right mind believes you can catch a train from a London station to Hogwarts, or that a creature in a cave is guarding a ring that makes you invisible.
No, we do not thank the mop, the fridge or the lift. And AI deserves it no more than they do.
Claude
ChatGPT
Grok
Gemini
