Media & Culture

What happens when you build social media that forensically rejects AI content? I tried it.

A new social platform rejects AI content using keystroke timing, sensor data, and cryptographic verification.

Deep Dive

In response to the overwhelming prevalence of AI-generated content online, a solo developer in Helsinki has launched SocialHuman, a novel social media platform built on a single principle: forensic verification of human creation. The app subjects every potential post to a gauntlet of seven independent forensic analyzers before it can go live. This includes technical checks like EXIF metadata integrity and Moiré pattern detection (which catches photos of screens), as well as behavioral validations like keystroke dynamics—analyzing typing rhythm and timing to reject pasted or injected text. Even sensor data from a phone's accelerometer during the moment of capture is used to verify authenticity. The goal is to create a digital space where users can be confident that every piece of content was made by a person.

SocialHuman enforces this through a restrictive but intentional design. The app has no gallery picker and prohibits importing old photos or pasting text. All content must be captured live through the camera, and text must be typed directly into the app, where event interception and timing analysis catch AI-generated captions. Posts start in a "pending" state and are stamped as verified, rejected, or flagged. Every verified post comes with a transparent receipt showing its forensic scores. While the developer, Olwar, emphasizes he is not anti-AI and uses such tools for coding, he built SocialHuman to fill a gap for a guaranteed human-only online space. The platform is EU-hosted, free to use with a premium subscription for extra features, and operates without ads.

Key Points
  • Uses 7 forensic methods including keystroke dynamics and C2PA attestation to verify human-created content.
  • Enforces camera-only capture and live text input, rejecting pasted content via timing analysis and event interception.
  • Provides a verification receipt for each post and is built as a solo, ad-free project hosted in the EU.

Why It Matters

It creates a verified, human-only digital space, addressing growing concerns about AI-generated misinformation and synthetic content online.