I think the scale argument is probably the strongest point here. The issue isn't necessarily that an AI-generated image or post is automatically bad, but that the barrier to producing hundreds of them is basically gone. A community built around human participation can get overwhelmed pretty quickly if the signal-to-noise ratio changes.
That said, I think there's a middle ground between "AI content is fine" and "AI content has no value." A lot depends on how it's used. Someone using AI as part of a creative process, then adding their own ideas, editing, or discussion around it, is a very different thing from dumping dozens of generated images with no context.
The same thing happened with low-effort content long before AI existed. Meme reposts, copied articles, generic fan theories, and promotional posts could also flood communities. The difference with AI is that the volume can scale much faster, so the moderation challenge is definitely different.
