"We wish it could have a better solution, but you all just couldn't behave" is often the argument in these debates. You seem to be dissatisfied with it, but tacitly accept it as the only way. Let me put a finger on why I will never agree with it:
1. The current situation isn't that apocalyptic.
2. The alternative is so much more horrifying.
The first is a weaker argument because there is always a bottomless bag for the opponents to reach into and grab counterexamples. But none of these things are new. It's been 30 years since the average person could freely communicate with anyone in the world over the internet. 20 or so since the maturing of online scams. Probably at least 15 since nation-level actors started using the internet for propaganda purposes. The elephant in the room is AI, but I feel like AI use has diminishing returns, much like in any other fields. It offers a quick boost, but it's not infinitely scalable. The internet can't be completely taken over by scams and garbage in a way where everyone just keeps consuming it and goes crazy in their own ways. It's almost an ecosystem - pour too much garbage in, and you'll see people retreating into private communities and self-enforce trust, no universal carding necessary. Right now it's just tech people doing this on a large scale, but at a certain point of misinformation and incomprehensibility, treating the internet as untrustworthy by default would just become a mainstream view. It would likely fracture into massive social networks with strong identity verification and small communities that enforce trust in their own ways. As much as I don't like most private businesses, but your suggestion that they could take upon the role of the verifier is more palatable to me because that still leaves air in the room for small, private communities, and because businesses don't have perfect communication with each other or the government, preventing the universal tracking of everything.
The second argument is the real one. Technology would enable governments to track absolutely everything about everyone if we let them. Now this is a space where AI actually provides a novel use case - if before you had to have suspicion about a specific person and have someone look into them, AI will enable them to inspect everything and then form suspicion or document undesirable behaviors by default. It will be like having a personal private investigator that sees everything you do, being on the lookout for anything unusual 24/7. Online identity verification is nothing like having a SIN or a driver's license, because those two are used in specific contexts and don't enable tracking-by-default. The internet is used for ALL communications, so automated tracking of it is tracking of everything we do and say. Instead of showing your driver's license to the officer one time in a defined spot, the officer would come to you because your driver's license is etched into your skin for all to see at all times, and they're concerned about your recent movements that were automatically recorded and processed by a billion cameras. It would be a world without secrets, a world of perfect knowledge and ruthless enforcement. I don't want to live in it. I'm fine with the current world, thank you.