What Is a Human-Only Social Network?

A human-only social network is a platform where every account is verified as a real, individual person, and stays verified. One human, one account. Checked on the way in, and checked again after.

That's the short answer. The longer answer is more interesting, because almost nobody agrees on what "verified" should mean, and the platforms making the loudest noise about bots are usually doing the least about them.

Why anyone is asking this now

Because people are no longer the majority online.

Thales and Imperva's 2026 Bad Bot Report found that bots generated 53% of all web traffic in 2025, the second year running that automated activity has outpaced humans. Bad bots specifically accounted for 40%, up from 37%, the seventh straight year of growth. Human traffic is now under half of everything happening on the internet.

Regulators have noticed. From 2 August 2026, Article 50 of the EU AI Act requires chatbots to identify themselves as chatbots and AI-generated content to be marked so it can be detected. When the law has to step in and say "tell people when it's a machine," the machine problem has stopped being theoretical.

So the question stops being philosophical and starts being practical. If you can't tell who's real, what would a network look like where you didn't have to guess? We wrote about how we got here in Dead Internet Theory and How Fake Is the Internet.

The three tests

A network is only human-only if it passes all three of these. Most platforms that use the phrase fail at least two.

Verified at the door. Every account proves a real person is behind it before they can participate. Not an email address. Not a payment method. A person.

Verified after. The check repeats. Accounts get sold, hacked, and handed over to software long after signup, so a one-time check has a shelf life measured in months.

One person, one account. Otherwise you've verified that humans exist, not that this account belongs to one. Account farms are run by real people. That's the point of them.

There's a fourth thing that isn't a technical test but decides whether the first three survive contact with real users: the community has to want it. If verification feels like a tax, people route around it. If it feels like a badge, they defend it.

What a human-only social network is not

A blue tick. Paid verification proves you own a working card. It says nothing about your pulse. Several platforms quietly turned identity verification into a subscription tier and hoped nobody would notice the difference.

A real-name policy. Facebook has required real names for twenty years and still removes fake accounts in enormous volume every quarter, by its own reporting. Asking for a name only filters people who are honest.

CAPTCHA. Machines now beat the tests designed to stop machines, and where they can't, solving farms staffed by underpaid humans fill the gap for fractions of a penny. CAPTCHA hasn't been a humanity check for years. It's a speed bump.

Moderation. Removing bots after they post is cleanup, not prevention. By then the content has been seen, the engagement has been counted, and the damage to your feed is done.

How human verification actually works

The technical field is called proof of personhood: verifying that an account belongs to one real, unique human, ideally without revealing which human. There are five broad approaches in use, and every single one involves a trade.

Biometric uniqueness. World ID, formerly Worldcoin, is the largest attempt, using iris scans taken on purpose-built Orb hardware to check nobody has registered twice. It works at scale. It has also been banned or restricted in several countries over the privacy implications of collecting iris data, and it requires you to physically visit a device.

Government documents. Systems like zkPassport read the chip in a biometric passport and use zero-knowledge proofs to confirm you're a real, unique adult without handing over your identity. Elegant, and useless to the enormous number of people who don't hold a chipped passport. The UN estimates over 100 million people have no travel documents at all.

Social graphs. BrightID and Circles infer personhood from relationships. If enough verified people vouch for you in credible ways, you're probably real. No biometrics, which is a genuine advantage, but newcomers with no connections find the door bolted from the inside.

Timed ceremonies. Idena runs synchronous validation sessions where everyone solves puzzles at the same moment, which makes running thousands of accounts expensive. It also means verification happens on the network's schedule, not yours, and puzzle-based tests get weaker every time a model gets better at puzzles.

Liveness and behaviour. On-device checks that confirm a living person is holding the phone right now: a moving face, a detected pulse, the involuntary tremor in a human finger. These are hard to fake, work on hardware people already own, and don't require anyone to scan their eyes. They need a phone, which is its own exclusion.

None of these is complete on its own. The serious systems stack several and accept that each one covers a different gap.

Why verifying once isn't enough

This is the part most platforms skip, and it's the part that matters.

Signup verification captures a single moment. Everything after that moment is assumed. Accounts are bought and sold in bulk, aged deliberately to look legitimate, then automated. Credential stuffing hands real accounts to whoever bought the leaked password. A person can pass every check at signup and hand the login to a script that afternoon.

Verification isn't a state you achieve. It's a habit the network has to keep.

Continuous verification means the checking never stops. Passive behavioural signals run in the background: typing rhythm, scroll patterns, the physical way a finger meets glass. Periodic challenges ask for something a bot can't produce on demand. Community accountability puts a real cost on vouching for a fake. A visible score reflects verified humanity over time instead of a checkbox from three years ago.

The difference is simple. Signup-only verification tells you an account was human once. Continuous verification tells you it's human now.

The costs nobody advertises

Anyone selling you a human-only network without mentioning these is selling you something.

Verification excludes people. Any phone-based system locks out anyone without a modern smartphone. Document-based systems lock out anyone without documents. Social vouching locks out anyone without an existing network. Every gate keeps out more than bots.

Biometrics make people uneasy, and they're right to be. The only defensible position is to process signals on the device, store nothing, and log a pass or fail rather than the raw data. Any platform hoarding biometric records is building a breach, not a network.

False positives punish real humans. Every threshold that catches more bots also catches more people having a bad day on a poor connection.

And verification costs money per user, which shapes what the business can be. Free platforms funded by engagement have a structural reason not to look too closely at where their engagement comes from.

So is a human-only social network actually possible?

Not perfectly. Nothing is.

The goal was never a mathematical guarantee. It's making automation cost more than it returns. A bot farm running ten thousand accounts works because each account costs almost nothing. Make each one require a face, a phone, a heartbeat, a vouch, and continuous behaviour that holds up over months, and the economics collapse. You don't have to make faking impossible. You have to make it not worth the effort.

That's the real definition. A human-only social network is one where pretending to be human is more expensive than just being one.

Where Hi Friction fits

We stack the layers rather than betting on one. Liveness and pulse at signup, passive behavioural signals whenever you're using the app, gamified challenges that verify you while you're having fun, community vouching with real reputational stakes, and a Human Score that reflects all of it over time. Posting is phone-only, so the proof is always available at the moment it matters.

Humans only. We checked. Reserve your username before the bots try.

Common questions

Is a human-only social network the same as a bot-free social network?

Roughly, though "bot-free" describes the outcome and "human-only" describes the rule. A bot-free platform might achieve it through moderation and luck. A human-only platform gets there by verifying everyone at the door and keeping it up.

Does verification mean giving up anonymity?

It shouldn't. Proof of personhood is specifically designed to confirm you're a unique human without revealing who you are. You can be pseudonymous on a human-only network. You just can't be a thousand people at once.

Can AI pass human verification?

Some of it, increasingly. Language models write convincing posts, and deepfakes handle video better every month. What they still can't do easily is produce a live pulse through a phone camera, replicate the involuntary tremor in a human hand, and keep both consistent across months of ordinary use. Layering is the answer precisely because any single check will eventually fall.

Do human-only social networks exist yet?

Verification infrastructure exists and is used widely. Full social networks built on it are still early, which is the honest answer and the reason we're building one.