Given there is "misleading or outright harmful" information generated by humans, why is it more pressing that we track such content generated by AI?

I suppose efficiency? It's easier to filter out petabytes of AI slop than to determine which human generated content is harmful