德国公共广播电视台播放了伪造的、AI生成的“ICE部队”逮捕移民家庭的视频。
German Public Broadcaster Ran Fake AI-Generated Clip Of 'ICE Troops' Arresting Migrant Family

原始链接: https://www.zerohedge.com/political/german-public-broadcaster-ran-fake-ai-generated-clip-ice-troops-arresting-migrant-family

德国公共广播电视台ZDF因其新闻节目《Heute Journal》播出人工智能生成的图像而受到批评,图像内容描绘了联邦警察逮捕一个移民家庭。该片段使用OpenAI的“Sora”制作(并清楚地标有水印),引发了公众愤怒,因为它将伪造内容呈现为真实。 ZDF承认了错误,并对图像未被正确标记表示遗憾,随后删除了该报道并用真实的影像进行了更新。该电视台声称,技术问题导致水印在传输过程中不可见。 这起事件引发了对ZDF使用人工智能生成内容以及过去是否曾使用未标记内容的问题的担忧,尤其考虑到该电视台获得了大量的公共资金。ZDF坚持其在使用人工智能方面的透明度承诺,但争议加剧了现有的偏见指控,特别是来自保守派团体。修改后的节目现在包含一份声明,承认了编辑变更。

相关文章

原文

Via Remix News,

The German public broadcaster ZDF has admitted to a significant editorial oversight after its flagship news program, Heute Journal, aired AI-generated images featuring U.S. Immigration and Customs Enforcement (ICE) agents arresting an immigrant family.

After uproar on social media over the fake news segment, which included a visible OpenAI’ “Sora” watermark on the screen, the broadcaster expressed regret over the error and has since updated the report to remove the synthetic content.

Critics pointed out that while it is becoming harder to differentiate fake AI content from real events, the appearance of the Sora watermark made it clear that this was AI content.

The controversy from the Feb. 15 report featured fake AI scenes of a woman and two children being led away by ICE. During the segment, ICE agents were referred to as “troops.”

When questioned about the incident, ZDF stated that the images should have been clearly marked. The broadcaster explained: “This marking was not transferred when the article was transferred for technical reasons.”

The question now is whether ZDF generated these images in-house. ZDF has declined to comment on whether the editorial staff was aware that the footage was AI-generated at the time of the initial broadcast.

If ZDF created them, the fact that arguably the biggest public broadcaster is creating AI-generated content for public broadcasting is raising concerns about how often AI-generated content has been produced without proper labeling in the past.

In response, ZDF reiterated its commitment to transparency, noting: “ZDF’s AI principles stipulate that AI-generated images are always transparently labeled.”

The incident caused further confusion when the original broadcast was temporarily removed from YouTube and the ZDF media library, leading some media outlets to report that the broadcaster had “deleted its fake video.” ZDF clarified that the removal was only a temporary measure while the editorial team replaced the AI sequences with authentic video and still images.

A revised version of the program is now available in the media library, accompanied by a disclaimer stating: “Video subsequently changed for editorial reasons.”

All German households are required to pay nearly €20 per month to fund ZDF and other public broadcasting outlets like ARD. That translates to billions every year.

The outlets are routinely accused of bias against conservatives, including negative reports targeting the Alternative for Germany (AfD) and a high rate of rejection for AfD guests on the networks .

Read more here...

Loading recommendations...

联系我们 contact @ memedata.com