What happened

On February 15, 2026, ZDF’s heute journal featured a report on U.S. deportation policies that included scenes of women and children being led away by emergency services. The segment also used a four-year-old clip of a child’s arrest in Florida (unrelated to immigration) as current evidence of enforcement actions. 

However, the emergency service footage was not real: it was revealed to have been generated using AI OpenAI's Sora AI video generator and had not been labelled by the broadcaster as generated using AI. 

Viewers quickly spotted the tell-tale indicators of fabrication and, following widespread criticism and social-media scrutiny, ZDF removed the fake footage from its online archive and revised the segment. 

Anne Gellinek, ZDF’s deputy editor‑in‑chief, publicly apologised and said the segment did not meet ZDF’s standards and should not have aired in that form.

At a time when AI-generated content is proliferating on social platforms, the footage was accused of distorting public understanding of a highly charged policy topic, undermining trust in journalism, and fueling political polarisation.

Why it happened

The incident was attributed to a breakdown in internal verification and technical metadata transfer. ZDF explained that while its editorial principles require the transparent labeling of AI material, the labels were "not transmitted" during the technical processing of the segment. 

The episode highlights that as AI-generated "slop" and propaganda flood social media, even high-resource public broadcasters are vulnerable to absorbing synthetic content into their reporting pipelines if robust, automated provenance checks (like C2PA) are bypassed or fail.