Example Scenario: How POY Verify Is Designed to Help a Social Platform Stop Deepfake Profiles

This example scenario follows a hypothetical professional social network where AI-generated profile photos are undermining trust. It shows how POY Verify is designed to let members prove a real human is behind their profile with an on-device liveness check, without the platform collecting biometric data from users in the EU or anywhere else.

This is an illustrative example scenario, not a customer case study. POY Verify is in private beta, so the company described here is hypothetical and no results are reported. It shows how the product is designed to work for this kind of platform.

The Scenario

Imagine a social platform built around professional networking and community discussion, with members in the EU and North America. Its reputation rests on authentic interactions. Then reports of suspicious profiles start rising, and review shows many use AI-generated photos to run spam, influence campaigns, and scams.

The Problem in Detail

In this scenario the deepfake accounts cause three connected problems:

• Trust Erosion - Members who discover they were talking to fake profiles start doubting every interaction, and engagement and retention suffer.
• Content Manipulation - Fake accounts amplify posts, steer discussion threads, and create the appearance of consensus that ranking systems then reward.
• Legal Exposure - Deepfake profiles that copy real people's likenesses create impersonation and fraud risk for the platform.

Traditional fixes run into a wall.

The platform needs to know a real human is behind a profile without becoming a store of biometric data.

The Solution

POY Verify is built around a simple idea: a platform needs the verification result, not the biometric data. In this scenario it works like this:

1. User triggers verification - A member taps a button to earn a Verified Human badge, which launches POY Verify on their phone.
2. On-device liveness check - The phone's depth camera, infrared emitter, and motion sensors confirm a living person is present, inside the Secure Enclave.
3. Cryptographic proof generated - The Secure Enclave produces a signed proof that a living human was detected. It contains no facial features or personal data.
4. Proof sent to platform - The platform checks the proof and adds the Verified Human badge to the profile.
5. No biometric data retained - Biometric data never leaves the device, so the platform has no biometric database to secure.

Why Zero-Data Helps with GDPR

GDPR's special category rules govern processing of biometric data. POY Verify processes biometrics only on the user's device and sends the platform a proof, not biometric data, which is designed to keep the platform's verification flow out of special category processing on its own servers.

The question to put to your own data protection officer: if biometric processing happens only on the member's device and the platform receives a signed proof, which GDPR obligations remain for the verification flow? POY Verify is designed so the answer is as few as possible, but each platform has to confirm it.

Before and After

Implementation Approach

A platform in this position would usually roll out in phases, starting voluntary and gradually making verification a stronger signal.

Phase 1: Voluntary Verification

Offer verification as an optional feature that earns a Verified Human badge. Power users, who have the most to gain from standing out from fake accounts, tend to verify first.

Phase 2: Trust Signal Integration

Fold verification status into trust and ranking: verified members' content and connection requests get a modest boost, which encourages verification without requiring it.

Phase 3: Flagged Account Review

Ask accounts flagged as likely synthetic to verify, or provide another proof of authenticity, within a set window. Accounts that cannot pass are restricted; real people who pass keep their profiles.

What to Measure

This scenario reports no results. A pilot would track the share of profiles flagged as synthetic, verification completion, member trust survey scores, impersonation reports, and moderation hours, comparing each before and after rollout.

The Data Privacy Advantage

The main design benefit is what the platform does not have to hold. With no biometric data on its servers, there are no facial images or templates to steal in a breach, and far less biometric data handling to document. Each platform should still confirm its remaining GDPR obligations with counsel.

Key Takeaways

• Deepfakes are a trust problem, not just a technical one - Fake profiles erode member trust and engagement, so the fix has to restore a visible trust signal.
• Biometric verification does not have to mean biometric collection - POY Verify processes biometrics on the device and sends only a cryptographic proof.
• Voluntary verification with incentives is a practical start - Badges and ranking signals encourage adoption before anything is required.
• Zero-data reduces breach exposure - With no biometric data on the platform's servers, a breach exposes proofs, not faces.