Commentary: The AI images in the "heute journal" are a serious breach of taboo

Eventually, it had to happen: a major media outlet is caught garnishing news with conveniently fitting AI images. The fact that we are now talking about the “heute journal” and the public broadcaster ZDF lends particular explosiveness to the whole affair. Even if it won't really surprise anyone who knows the peculiar mix of elitist self-image and sluggish error culture that is typical of public broadcasters.

The fact is: Someone at ZDF thought it was a good idea to illustrate a report on the sometimes brutal and deadly operations of the US border police ICE in the “heute journal.” Not only with an AI video because crying mothers and children fit so nicely, but also with another sequence that, while not faked, is old and irrelevant to the topic.

Red line crossed

With this, ZDF – and one must say it so clearly – has crossed a red line. Faked images in the news environment are a serious breach of taboo. They distort the reality that journalists are supposed to portray. They contradict the principle of truthfulness, which is the basis of journalism. Furthermore, they destroy the trust that is the foundation of the relationship between media and their viewers, listeners, and readers.

An opinion by Volker Briegleb

In his 20 years in the newsroom at heise online, Volker Briegleb has seen many trends come and go. He's not sure whether the internet or Hertha BSC is the biggest disappointment of his life.

As if that weren't bad enough, ZDF gets tangled up in contradictions when the first questions arise. A "technical error" during the transfer is blamed for the AI scene not being labeled as such. The broadcaster removes the report from the media library. When it goes back online, the scene is not labeled as AI, but simply deleted and replaced. And the other one too.

ZDF does not answer questions about the exact procedure and responsibility for the broadcast report. Nor does the broadcaster initially respond to the obvious question of why the editorial team apparently deliberately incorporated an AI sequence into a report and whether this complies with internal rules on AI use (Spoiler: it doesn't), apart from attempting to retroactively reinterpret the report's topic.

Then a half-hearted turn: On Tuesday, ZDF admits the mistake, but continues to waffle about missing labeling: "According to ZDF's rules, this sequence should not have been used without labeling and without context." In the evening, Deputy Editor-in-Chief and Head of News Anne Gellinek appears on camera and manages an apology.

In the subsequent report "in our own interest," ZDF finally admits, referring to internal rules: "The use of AI-generated images and videos of people, events, and political contexts is not possible in the news sector." In plain text: The editorial team should not have cut the video generated with Sora into the report in the first place – not even if it is "labeled." In a report from Wednesday, ZDF explains that the lack of diligence by the "author" and insufficient review by the editorial team are responsible. The corrected version of the report has since also disappeared from the media library.

Stupidity or Intent?

Which brings us to the question of how all this could have happened – and to "Hanlon's Razor." This is an English proverb that states: "Do not attribute to malice that which is sufficiently explained by stupidity." Unfortunately for ZDF, there are now further findings that allow us to largely rule out stupidity.

Because ZDF had already broadcast a slightly longer edited version of the report by chief correspondent Nicola Albrecht on February 13th in the "Mittagsmagazin" (from 01:20:00). At the questionable point with the authentic images, which were then also visible again in the corrected version. No AI, no false context. This also makes it clear: The narration was also refined for the "heute journal."

However, the official ZDF version places the responsible person in the New York office. ZDF does not explain why the author would deliver an impeccable report for the "Mittagsmagazin" and then falsify it herself for the "heute journal." Beyond its written statements, the broadcaster does not answer any questions.

Thus, the suspicion arises that the editorial team of the "heute journal" subsequently manipulated the report by the New York correspondent because the real images do not support the narrative clearly enough or are not emotional enough.

That would bring us to "fake news" and "disinformation,” the fight against which public broadcasters are currently championing. Or to put it very clearly: Shabby propaganda spread by a public broadcaster. ZDF must now transparently address this and draw real consequences that go beyond the announced training for employees – a "Sorry, folks” is not enough.

(vbr)