(Originally posted on LinkedIn.)
OpenAI’s ChatGPT is not intelligent. It is distilling knowledge fed into it by humans mining billions of data points. It’s a glorified chess computer.
I am deeply concerned that this will increase the amount of misinformation floating around, because it has the confidence programmed into it by the very people who should know better. Speaking in absolutes is dangerous.
I am also deeply concerned that it will tank content creation industries, which are already struggling with misappropriation, and little or no attribution. Artists are still reeling from the effects of Google Image Search, not to mention “AI” generative art, which is remixing existing works without attribution (or exchanging money).
Educational institutions and publishers will have a much harder time identifying plagiarism (and to be very clear, any content that tools like ChatGPT “generate” is taken from other sources, because this is not an intelligent or creative system).
Is it impressive? Sure. Is it smart? No. Is it practical? Also no. Will it take jobs away from people? Probably, but at a not insignificant cost.
I’m sad because of the major ethical issues this raises. I’m also sad because we keep saying Artificial Intelligence. There’s nothing intelligent about it, artificial or not. The only magic here is the magical thinking involved.
If you want cool art, pay a cool artist. If you want to have a conversation, speak to a friend or loved one. Critical thinking is the one superpower humans have left.