AI is getting frighteningly good at pretending to be us. One company thinks your eyeballs might be the solution.
Remember when proving you were human meant clicking every square containing a traffic light?
Good times.
Because apparently the robots got better.
Automated bots now generate more than half of global web traffic. AI agents are getting good enough to browse websites, shop, book things, message people and generally cosplay as us online.
Which creates a slightly ridiculous problem:
How does the internet know you're actually human?
Sam Altman has an answer.
Your eyeballs.
Why you should care
Because “prove you’re human” could become the internet’s next password.
As AI gets better at impersonating us, your face, eyes, or other biometrics could become the ticket into apps, dating, shopping — even banking.
What you can do
Treat biometrics like a password you can’t change. Before scanning, check who keeps it, for how long, and whether you can delete it.
👁️ Meet the Orb
A company he co-founded, Tools for Humanity, built something called the Orb.
And yes, it looks exactly like something a movie villain would ask you to stare into.
The Orb takes high-resolution images of your eyes and face, checks that you're a unique human, then gives you a World ID — essentially a digital “yep, real person” badge. World says the images are transferred to your device and deleted from the Orb after verification.
You only need the eyeball scan once.
After that, websites and apps could ask for your World ID instead of making you identify another godforsaken bicycle.
And this isn't some tiny experiment.
Around 18 million people have already completed Orb-level verification, according to a World executive interviewed by The New Yorker. World ID is already being pushed into things like dating, shopping and online services.
But why would anyone want this?
Think about the internet we're walking into.
So eventually, being human itself becomes scarce information.
That's weird.
For most of internet history, machines had to prove they weren't pretending to be us.
Now we may have to prove we aren't them.
🍌 The Banana Peel
There is, naturally, a tiny issue.
To solve the problem of giant technology companies creating increasingly convincing artificial humans…
we may end up giving another technology company infrastructure for verifying actual humans.
World argues its system is designed around privacy and doesn't require your name or email. But regulators in several countries have investigated or restricted the project over privacy, biometric-data and consumer-protection concerns.
So this isn't really a story about a shiny eyeball scanner.
It's about something much bigger.
The internet used to ask:
“Are you a robot?”
AI is making the next question:
“Can you prove you're human?”
And humanity somehow made CAPTCHA lore important.
The monkey misses the traffic lights already ;(
