Last month’s shit-show was OpenAI’s new ChatGPT Agent nonchalantly clicking through Cloudflare’s “I am not a robot” box, narrating the process like a teacher. Like oh, this step is necessary to make sure I’m not a bot. It was absurd. The screenshots went viral. Reddit went crazy. Hacker News lit up with hundreds of comments. The reason everyone’s freaking out is because those screenshots mark the collapse of the whole concept of proving you’re human, or what we used to call “proof of humanity.”

The early internet had the basic idea that if you were a real person, you could prove it. That was the old deal. You’d type the squiggly letters on a CAPTCHA, click all the fire hydrants, upload a pic of your driver’s license, maybe even hop on a short video call and say yeah it’s me. Those worked because humans were unique and machines couldn’t really fake that uniqueness.

Well the internet doesn’t work like that anymore.

AI can write like you. It can talk like you. It can generate a profile pic that looks like a real selfie with background clutter and all that. It can invent years of fake posts and fake life events in seconds. It can even mimic the small mistakes that used to be the giveaway signs of being human. And now it has captured CAPTCHA.

And while the story about ChatGPT clicking the “I am not a robot” box is surreal, a script could’ve done that a decade ago. The difference is that now the bots don’t need to hide. Years ago, automated systems would try to pretend they were human to slip past human‑check mechanisms. Now, with recognition systems, they can simply be recognized by the system as valid actors, whether they’re human or not.

Now that AI can do the things only a human could do better than a human, the standard has changed to recognition status. The old approach that asked, “Are you a person?” has switched to a new question, “Are you recognized?”

Recognition means you’re in the database. You have a valid account tied to the right government ID or trusted login token. Whatever the system uses, it already knows you.

If you’re outside those systems, it doesn’t matter if you’re not a bot. The shift from human vs. bot to recognized vs. unrecognized changes the stakes. It’s about who gets to exist online, which also means who gets to exist in modern life.

For example, say you send in your résumé for a job. The HR’s applicant-tracking software scans it but you’ve been flagged before, maybe an account you used in the past got linked to spam or your name is similar to someone banned or something like this. You’re not in their whitelist. So you don’t get an interview, because the system doesn’t recognize you.

Banking is another one. Say your account gets locked for “suspicious activity.” So you walk into the bank with your ID but the bank now relies on automated KYC verification. If your face scan fails or your address doesn’t match perfectly, you can’t get your own money. The teller might even know you, but she has to go with the machine because she doesn’t want to get in trouble. You can eventually get it sorted out but only by going through the system’s full re‑verification process to be brought back into alignment.

What’s so absurd about all of this is that the very thing that killed proof of humanity is now deciding who counts as “real.” And neither the AI nor the infrastructure behind it cares whether you’re human. It only cares whether you’re aligned with the system.