Профиль

"AI" is not intelligent. Chatbots don't *understand* the question you ask, or the text they produce. If we called it what it is, e.g. "probabilistically produced synthetic text" (Bender & Inie), would as many writers be keen to let it loose on their manuscripts, or to let it "write" them? 🤔

"Unlike other non-living things, AI systems can form relationships with humans." My sweet summer child. Humans can *feel like* they've formed relationships with AI systems. As they can with many non-living things (e.g. one woman married the Eiffel Tower) www.theguardian.com/technology/2...

Could AI be conscious?Experts believe it’s at least possible. We urgently need a plan to navigate the ethical implicationswww.theguardian.com

In fact, one of the first suggestions is not to use the term "artificial intelligence", but the more accurate "probabilistic automation". If that became the accepted term, I wonder if as many writers would be keen to let it loose on their manuscripts. Or let it generate their manuscripts 🤔

Emily Bender and Nanna Inie: "all LLM output is probabilistically produced synthetic text". AI is not "intelligent". The chatbot does not have a clue what it just said to you. It does not *understand* the text it generates. So here's how to talk about it without making it sound like a living being