My Brazilian eye surgeon told me that 'they' (the authorities) program mistakes into ChatGPT (and other free AI platforms) so that we feel as though it doesn't work well. I find this idea a bit strange, but who knows? Perhaps it's true. My brother told me AI makes mistakes with simple mathematical tasks, and I thought this was very strange as calculators never make mistakes. AI can't use apostrophes either, which seems silly to me. Apostrophes are easy.