Onchain Atlas

BrightID Sybil Resistance

Proves 'this is one real, unique human' by checking they show up in a web of verified in-person or video social connections, without using ID documents.

Also called: social-graph identity verification · BrightID uniqueness verification

What it is

BrightID-style sybil resistance verifies that an account belongs to a real, unique person by analyzing a social graph of connections formed through verification parties (small group video calls or meetups) rather than by checking a government ID or biometric. Other applications (grant programs, quadratic funding rounds, DAOs) can then query "is this a unique human?" without handling any personal data themselves.

How it works

  1. A user creates a BrightID (or similar) identity and connects it to other users they know, forming edges in a shared social graph.
  2. Periodically, users join small-group verification sessions (video calls) with strangers, where participants mutually vouch that everyone present appears to be a distinct real person.
  3. Graph analysis algorithms examine the resulting network structure — looking for patterns like isolated clusters, unnaturally dense connections, or accounts with no organic ties — to flag likely sybil (fake/duplicate) identities.
  4. Accounts that pass analysis receive a "verified unique human" status, often expressed as a signed attestation the user can present elsewhere.
  5. A third-party app (a quadratic funding round, an airdrop, a DAO) queries or requires this attestation before letting an account participate in ways that assume one-person-one-vote or one-person-one-claim.
  6. Because verification is probabilistic (graph-based), it's periodically re-run and status can be revoked if later analysis finds an account's context looks sybil-like.

Why designers use it

  • Enables "one person, one vote/claim" mechanisms (quadratic funding, fair airdrops, one-account-per-human governance) that would otherwise be trivially gamed by an actor creating many wallets.
  • Avoids collecting government ID or biometric data, which most crypto-native users and privacy-conscious protocols want to avoid holding.
  • Reusable across many applications: get verified once, use the attestation anywhere that accepts it, rather than re-verifying per app.
  • Doesn't depend on a single centralized authority deciding who's "real" — the graph structure itself does the work.

Failure modes

  • Small, effective sybil rings can still form if a group of colluding fake accounts vouch for each other convincingly enough to fool the graph analysis, especially as attackers learn the detection heuristics.
  • Verification parties require real-time coordination (scheduling a video call with strangers), which creates friction and excludes legitimate users who can't participate — hurting inclusivity precisely where the mechanism is trying to be fair.
  • The graph-analysis approach doesn't scale trivially; as the network grows, false positive/negative rates and the cost of manual review both grow.
  • Reliance on a single graph/verification network creates a chokepoint — if that network is compromised, unavailable, or centralizes around a small set of validators, every downstream app inherits that risk.
  • Legitimate users with unusual but genuine social graphs (very online, new to a region, isolated) can be incorrectly flagged as sybils, denying them legitimate participation.

What to check before using it

  • What's the actual false-positive and false-negative rate for the sybil detection in practice, and is that acceptable for the stakes involved (a small airdrop vs. a large treasury vote)?
  • Does the verification process create meaningful friction that excludes real users who should be included?
  • Is the underlying social graph/attestation network itself decentralized enough not to become a single point of failure or capture?
  • How often is verification status re-checked, and can compromised or sybil accounts that slip through get cleaned up later?
  • What's the cost (in time, coordination, or privacy trade-off) a real user has to pay to get verified, relative to the value being gated behind it?

Experiments that used it · 2

Shown oldest first, so you can watch the design evolve.

clr.fund
A permissionless, trust-minimized quadratic funding protocol for Ethereum public goods that used MACI zk-SNARK voting to make contributions bribery-resistant.
2020 partial success
1Hive Gardens
1Hive's Gardens turned the 1Hive DAO's own operating system — conviction voting, a community covenant, and the Celeste dispute court — into a reusable template for bottom-up, continuously funded DAOs.
2021 partial success