Large Language Models – the platforms which power chatbots like ChatGPT – are basically joining dots in enormous datasets of text, and guessing the next word in a sequence, he said. He added that the existential threat warnings issued by some AI experts relate to systems which don’t yet exist.
