I think this is a nice gesture but I also think it's way too charitable for the Anti-GenAI people currently complaining. Their issue isn't that they want to avoid it, they want to stamp it out, they want it gone. They're not going to hide posts or keywords, they're going to brigade them. They already brigade and harass explicitly GenAI communities and don't block them. They've harassed me multiple times, told me to die in DMs and even impersonated me for running such communities. Tagging isn't going to help, they're just going to use it to hunt down people who post it and brigade the posts or target the users, because they aren't angry that they are seeing it, they're angry that it exists and they wish to stamp it out, no matter the cost.
In short this is a good solution with good intentions but it assumes a level of good faith that just isn't there. I'd agree with this if the problem was really people just not wanting to see it, but the problem is much deeper. I'm sure that once people start doing it, the goal posts will be moved and they'll just stop using "untagged" as a reason for complaining.
I think that the AI friendly nature of our instance needs to be clearly written in our instance sidebar (@[email protected] and @[email protected] can help with that) which would help out since the instance sidebar is visible in all dbzer0 communities which are remote. It should probably also be included in the application prompt as a disclaimer, since I'm seeing a lot of dbzer0 people who signed up and are complaining about AI (guys what the fuck are you doing here on an AI instance if you hate AI).
Also moderation, moderation is the only real way to deal with the Anti-AI brigading and trolling. Preventative moderation can also be used as well if things get bad enough and/or don't improve.