Detect Deepfakesby Resemble AI
Deepfake case study · Image

Debunked: Pictures of Ursula von der Leyen with Jeffrey…

Viral AI-generated images falsely depicting European Commission President Ursula von der Leyen with Jeffrey Epstein have been debunked using digital…

Incident date
Sep 2024
Target
Ursula von der Leyen
Updated Sep 6, 2026 · 1 min read

In early September 2024, social media platforms were flooded with AI-generated images that falsely depicted European Commission President Ursula von der Leyen in conversation with the late convicted child trafficker Jeffrey Epstein. These fabricated photos, which circulated widely across Facebook, Instagram, TikTok, and X, were used to fuel baseless conspiracy theories linking political figures to Epstein’s criminal activities.

What happened

The viral images purported to show Epstein whispering into the ear of von der Leyen at a formal black-tie event. To add a veneer of authenticity, the creators embedded fake metadata in the corners of the photos, including "dossier numbers" and "document numbers" referencing the year 2012, alongside the French label "Fichiers Epstein" (Epstein files).

The deception was confirmed as artificial through technical analysis. Investigators utilized a tool provided by OpenAI to detect a "SynthID" watermark—an invisible, digital signature embedded by specific AI image generators to identify machine-created content. Despite this definitive proof of fabrication, the images continued to gain significant traction, with one post on X accumulating over 244,000 views. These posts often included inflammatory, evidence-free claims about the political affiliations of von der Leyen and other public figures. This incident follows a pattern of similar disinformation efforts, including earlier attempts this year to circulate AI-generated images falsely suggesting that Jeffrey Epstein was still alive and living in Israel.

Sources