This 'College Protester' Isn't Real. It's an AI-Powered Undercover Bot for Cops

I sometimes wonder whether the end result of this proliferation of bots is the creation of a "premium" Internet where you are authenticated as a real person before entering. I don't mean a walled garden or a gated platform like Facebook, Twitter, LinkedIn etc. I mean some sort of application layer protocol running on top of TCP that has real world authentication built in. Any application built on top of that protocol is guaranteed to be used by only real human beings.

> Any application built on top of that protocol is guaranteed to be used by only real human beings.

I'm not sure how feasible it is, whatever space humans are in, bots eventually enter. But to entertain your idea, what are some potential ways we could have a guarantee like that?

The thing is, services should exist on this internet, and this means pollution by bots is inevitable.

Sam Altman is trying to build this with the world app and retina scanning. Who knows if it will go anywhere.

I've been thinking about this. I think we may have enough standards to build it today on top of existing protocols.

So what happens if Eve authenticates as a human, then gives Mal (an AI) access to her screen and keyboard?

Humans have shown them selves to be more than willing to post generated content on behalf of a bot. And worse, bot farms often employ a bunch of humans to mass post bot created content.

Anything that can prove that a human is sitting at a keyboard on the other side of the world can be fooled by a sufficiently advanced bot.

What set of standards could possibly facilitate this?

Yes, see other places where humans are supposed to be guaranteed, and bots get banned--like online poker or video games.

They're crowded with bots.

Then busting "Mal" also busts Eve

Eve is vouching for Mal. And when Mal1, Mal2, Mal3 etc are all vouched for by Eve, then they are all trivially linked.

The far bigger problem is "how do you vouch for a single entitiy". How do you prevent Eve having multiple unlinked accounts.

> build this with the world app and retina scanning

So how would the interface/UX work with that, for each outgoing request you'd need to scan your retina, so you'd have something like a Yubikey at home, but with a little retina-camera instead?

I'm afraid it's one of those things where everyone thinks it would be nice if it existed but the real demand is not strong enough to support an implementation. But who knows, maybe it will change someday.

Eve was desperate for grocery money, made a couple bucks this way, and is now permanently locked out of society for it. Now what?

I think an individual could have multiple digital identities, with varying levels of trust in the veracity of that identity. They would have the ability to create/update/delete (revoke) those identities.

Is it not? Maybe not the full tech level, but real ID requirement for facebook is already a thing for a while.

* scans eye *

* pushes button on keyboard-button-pressing machine *

The analog hole works just as well in reverse.

Nobody cares because we're in a hellish dystopia where we're taught (and required) not to! Remember, citizen, empathy is evil!

I hope it will actually result in more real-life engagement throughout society once most people realize personas on the internet are for the most fake. Also -- disappearing messages by default!

Initial verification requires you to physically travel to the nearest verification center, and present yourself in person.

In the context of this post, it's bigger than that. Eve would be a government employee who can likely make new "human" profiles for Mal instead of using her own. Any centralized "human" identification system so we can vouch for a single entity-identity would probably be either run by the government or indirectly controlled by it (someone has to be able to input birth/death records at least, and countries like China aren't going to let e.g. Worldcoin operate within their borders unless they can control accounts).

Spam bots would use stolen/purchased credentials and get shut down. State-level bots would be free.

I don't think Facebook satisfies the demand for interaction with actual human beings.

In-person visits are of limited value when shady middlemen can give poor folks a fake ID and pay them a pittance to register as "you."

Me: "Hello I am from Autoritarianistan, I would like your premium internet in our country."

You: "Cool, we'll get tons of infrastructure in your country and make lots of money because you'll force everyone on it.

Me: "Hey, this is working out great. Bring your team over to Auth'istan for a business trip it will be great.

You: [Partying in said country]

Me: (to you) "Come over to this dark room a minute"

You: "Eh, this room is kinda sketch and why do those guys have a hammer and pliers"

Me: "So there is an easy way or hard way to this. Easy way, you give us unfettered access for our bots to spread propaganda on your premium internet. Hard way, we toss you off this building then we'll throw the rest of your team in a dark prison cell for doing the cocaine on the buffet table. Eventually one of them will give in and give us access."

Me: "So which is it going to be"

The $5 wrench is typically the winner.

https://m.xkcd.com/538/

Well, shit, we found the guy that would build the Torment Nexus from classic sci-fi novel Don't Create The Torment Nexus.

Most countries don't even have accurate population figures, let alone an accurate list of who is and isn't on there, even if you assumed they wanted to operate openly with no corruption and perfect intentions.

The only thing you could do is build an identity from scratch, karma as you will, but then if it was actually valuable people would sell that (see rich people buying high-powered gaming accounts for the lolz etc)

Hah good point. Which leads to what the other comments brought up, what's to stop real people from selling their ID to bot posters?

I can imagine a system which assures one ID per actual verified meat human. The ID can be banned from communities, blocked by individuals or kicked out of the platform.

I think this would limit the supply of IDs, so a big improvement over the current situation where bot posters can just farm fake IDs. Obviously this has some assumptions - for example that most humans joining the platform do so in good faith and not to just sell their IDs to bots.

The idea of "disappearing messages" is bullcrap. Anyone can just modify their client to store them.