At the article level, it is indeed mostly harmless, except when they catch some Google bias such as removing dates and edit notices from blog posts (because evergreen articles supposedly rank better), force a certain word count, H1/H2 count and paragraph size (which is why you get a life story before a recipe), or when it teaches you to disguise keyword stuffing and you get articles with semi-nonsensical paragraphs that can rank high. This is mostly annoying, though.
But IMO the biggest problem is SEO at a strategic level, which is what's done with the help of tools like Moz and Ahrefs. Websites often hire content agencies to make a large number of low-quality articles targeting popular searches that don't show too many results yet. Basically "balancing" demand/supply. But this practice puts non-SEO-optimized results at the bottom and replaces them with low-quality articles written by freelancers. This forces non-commercial publishers into bigger websites (like StackOverflow, Quora or Reddit). It also causes a lot of duplicate content, since those content agencies uses existing content from results as the basis for their own articles, only SEO optimized. The articles are of course rewritten so it's not word-by-word plagiarism. They also help evading Google measures to fix results. The latest trend when I was still in the industry was having unrelated companies helping each other do inorganic link-building by putting random links in more popular websites also controlled by those agencies, but that's probably old stuff by now.