Why a Verified Human Crowd Matters

A preference test is only as good as the people taking it. If a result can be swayed by bots, scripts, or one motivated person voting a thousand times, the number is worthless. IABT’s answer is a verified human crowd: every vote comes from a real, unique person, and each person counts once per test.

The two failure modes we design against

• Bots and automation. Cheap to spin up, easy to point at a survey. A handful of scripts can drown out genuine opinion.
• Ballot stuffing (Sybil attacks). One person creating many identities to vote repeatedly for the outcome they want.

Ordinary logins don’t stop either problem — email addresses and accounts are free and unlimited. IABT layers stronger checks on top.

How verification works

IABT combines several signals, strongest first:

1. Proof of personhood — World ID. Proves you’re a unique human without revealing who you are. Each test action produces a privacy-preserving nullifier that lets IABT count you exactly once — and reject a second attempt from the same person — without ever linking your votes to your identity.
2. Verified identity — ID.me. For tests that need verified attributes (age bracket, region, profession), ID.me supplies them with the user’s consent, so results can be segmented by real demographics.
3. Liveness — self-hosted vision. A lightweight, self-hosted check confirms a live person is present, not a replayed recording. No biometric images are stored — only a pass/fail.

Signing in and proving personhood happen in the same step: your World ID verification both establishes your session and ties each verdict to a verified, unique person.

Privacy by design

Verifying that you’re a unique human is not the same as knowing who you are. IABT keeps these separate:

• World ID nullifiers are per-action and cannot be correlated back to a person or across tests.
• Liveness checks store only a yes/no result — never raw frames.
• Identity attributes from ID.me are used for segmentation only, with consent.

The goal is a large, diverse, provably human panel whose collective judgment you can trust — recruited at the scale of the open internet, but without the noise that usually comes with it.

Ready to lend your verdict? Here’s how to participate.