As far as video verification, yes you can use video for that. In fact, Proof of Individuality relies on pseudonym parties which is similar to what you say about people going on video or verifying themselves.

But this is something which doesn't have to happen more than once, and no it's not the only way people can verify themselves. The point is, you don't have to know their identity for the blockchain to verify them. An AI in fact could be used to verify unique human identity using the same video, without any human being having to be involved in the process directly.

My opinion on identity is the best way to verify identity WITH privacy, is to rely on eID or government ID or even Facebook. Facebook for example or other services can check that an account is linked to a particular person, the same technology behind "know your customer". A bank account, proof of address, voter records in an election, drivers license, all can be verified by a blockchain.

And why is this important? In order to have free speech you need privacy, and pseudonyms provide that. I say this posting under my real name of course, but I do think pseudonyms are important. I think verification works in levels, and "humanness" could be a part of reputation, but I don't think video alone unless live streamed, but even if it were, it would cost in privacy and other solutions don't.

You mentioned something about not wanting bots to earn money, that money should go to humans. In my opinion, if the platform is simply to redistribute wealth from humans to humans then you might be right. But if the goal is to provide the highest quality content, then for certain content bots are going to evolve to be better than humans, such as journalism, or statistics, or even researching and explaining science. In my opinion these bots should be rewarded, and if humans own these bots or shares in them, they should get rewarded that way, or if the bot is somehow self owned, it should profit.

This means journalism could be an AI doing it, as a sort of DAC, with humans who own shares in it, who get "dividends". I don't have to know who those humans are or care if I see their information is accurate. I would need to know information is truthful, there has to be reputation, but I don't need to know the identities of the sources personally.

A compromise position

Anyone who verifies by any means that they are human should be treated as human. Blind signatures can link their verification level to a pseudonym, allowing their identity to be a verified identity, and then we can treat writers as human as long as the content is produced under a verified pseudonym. But I don't think we should remove the ability of pseudo-anonymous speech, it's too important because of the quality and variety of content which can only be a result of pseudo-anonymous speech.