Generative AI also needs improvements regarding the generation of content. One time, I saw that some of my classmates were using ChatGPT to solve their homework. I don't know if they checked if the information it gave was correct, but if they didn't, then I wouldn't be surprised if the teachers saw errors in its text.
And one day, when we were preparing ourselves to study a kind-of long literary work for a test, one of my classmates made a summary of it, which was apparently made with AI. When I read it, I was surprised at how the text was so inaccurate, as most of its content didn't coincide with what the literary work really said. I had to point him out that there's literally a website that had the work already summarized correctly.