I hate that “AIs” like ChatGPT (aka people-pleasing hallucinating plagiarism engines) are trained on work stolen from writers, including every single one of my books. I hate more that it’s eroding our ability or even desire to think.
I get it. Thinking is hard. It’s ALSO the work. It’s the whole point. It’s what makes us human.
We’re outsourcing the most vital human thing, our ability to reason, think, imagine, question, the pleasure of figuring something out, working out how we feel, to the most mediocre simulacrum of our wild and weird and magnificent brains, that is ALSO killing the planet.
LLMS use more electricity, more water than whole cities to answer the most basic bitch questions, to write paint-by-numbers college essays, to write dull copy and remix poetry, or worse, a text message to your spouse or an email to your boss or a customer complaint form.
Also, as the tech gets more powerful, it’s becoming more unreliable, hallucinating even worse and just wrong answers. (As my neuroscientist friend Dr Anil Seth explained to me, it’s more like the confabulations of your aging relative with dementia – reaching for an answer that sounds almost right to make you happy, rather than LSD fractals and pink elephants)
I did a talk about this recently, for The Turing Institute’s AI UK conference on the Imagined Futures panel. I’m so angry and frustrated with the current state of “AI” that I did a drastic, and impossible re-imagining of a different kind of artificial intelligence that makes us MORE human not less. (Spoilers: it riffs off the South African philosophy of ubuntu and also the interconnectedness of mycellium networks).