You could have something that is more akin to "verified-human" rather than verified identity. Though that alone would not prevent actors from posting llm content from verified accounts. So then you could have an additional layer where an account can lose its "verified-human" status if it posts llm generated content without disclosing it as such.

▲

augment_me 3 days ago | parent [-]

What's the incentive for such a system every existing? Let's say you are ANY current actor, like Google, Cloudflare, Meta, X, etc.

Why on earth would you give up the possibility to track people by having a "verified human" level, and spend a bunch of resources on validation, moderation, etc?

There is 0 incentive

	▲	ddj231 3 days ago \| parent [-]
		it doesn't necessarily have to be web-wide. it could be a single social space which is being bombarded by llm content and is incentivized to show users what is or is not verified without losing the informal culture of usernames vs. actual names. I'd say its not all too different from what X did. But there are different technologies at play now. In any case the actual content of your text these days acts already as a kind of signature of llm vs. human made. So this is just codifying it.