As soon as there's any discriminatory action taken based on that identification there's the incentive to lie and not use it and it becomes the same cat and mouse spam/antispam detection rat race sites already face.

But how would you enforce that, particularly against bad actors in countries that care less about enforcing laws and norms?

Spam calls are still a massive scourge even with Caller ID, SHAKE/STIR, and not for lack of trying.

Micropayments

Or proof of work to at least deter the increase in transfer volume. Neither are environmentally friendly.

The solution is probably legal, not technological.

We need a legal solution to automated spam, AI or not. It does tangible damage to communication systems when 99/100 of your e-mails doesn't come from a human being.

What about spam from other countries? Residential botnets?

That's a problem for whoever gets in charge of solving spam to solve. Not me.

If humanity's "solution" to AI is to go to the government of all institutions - the very thing we have hoped for the internet to overcome, for a century before its invention - then we'll truly have fallen into a poetic trap.

I've never hoped for that. The opposite of the government owning the internet is cloudflare owning the internet.

The opposite of one large unaccountable entity owning the internet is not a different one doing so.

To suppose that this next phase of human evolution can be "owned" at all is a bit of a stretch. These behemoths can threaten us with violence, and carry out those threats, but they can't stop radio waves or the use of language as expression. They can't remove knowledge form the universe or change change the laws of mathematics.