I've always been skeptical of generative AI being used to make slides. I like the idea of it, and maybe someday we will get there. But right now, they all feel like they are trying to fully automate the process where what I really want them to do is automate the tedious parts, co-piloting with me.

Because what ends up happening is I need a high degree of precision and I have a very concrete idea of what I need to present on a slide and exactly how to do it. But because it's fully automated, I have to instead convince the AI to generate my concrete idea on its own. Then I just don't use it entirely. Meanwhile, I go back to tediously making little shapes and moving them into precise locations, aligning things, and so on.