Particle.news
Download on the App Store

ZDF Apologizes On Air for AI Clip and Misdated Video in ‘heute journal’

The broadcaster says it will tighten verification with mandatory training after breaching its own AI rules in a report on U.S. ICE operations.

Overview

  • ZDF confirmed that a Feb. 15 segment on ICE used an AI‑generated video bearing a visible Sora watermark and a separate 2022 Florida arrest clip presented as if current ICE footage.
  • The network removed or re‑edited the item in its Mediathek and then aired a detailed correction, with news chief Anne Gellinek apologizing on air and calling it a “double error.”
  • ZDF stated that its guidelines prohibit AI‑generated images of people or political events in news unless the piece explicitly examines AI fakes, acknowledging a label would not have sufficed.
  • Following internal review announcements, a spokesperson said the broadcaster will impose stricter video verification procedures and require extra training for staff.
  • Politicians and media critics condemned the lapse and demanded full clarification, intensifying scrutiny of how public broadcasters verify social‑media footage and detect deepfakes.