As someone that has messed with Stable Diffusion for generating images, this is a non issue. AI images are challenging to prompt well and generate results. Good results are the result of clever prompts and what really amounts to a new form of artistry. The hype over AI is mostly due to mainstream garbage hype media and people that can not distinguish between technology and magic. If anyone opens Stable Diffusion and tries to just write a prompt, the output is usually garbage. If they are super lucky, the first result may work, but it will go downhill from there. Good prompts take practice, research, and lots of extra software tools. It is a skill all its own. If you are going to restrict AI for anything, it should be limiting it to a tag in the title, must include the entire prompt and seed, and the positive prompt must match the image reasonably well. This would limit post submissions to real artistry and it would enable anyone that is interested in this technology to reproduce the results and explore them more.