Advertisers Challenged By TikTok Deepfakes Of Hamas Victims

In a video image posted on TikTok, an Israeli woman shared her experience of the Hamas attack on Israel, but it is actually not the 22-year-old German-Israeli citizen Shani Louk who is describing the event. She was killed in the Hamas siege during a music festival on October 7.

This is a deepfake created by artificial intelligence (AI) of the woman’s image posted to TikTok.

Someone had used publicly available pictures of Louk to make the video, according to Bloomberg, which reported seeing dozens of these types of videos in numerous languages across the social platform. CNN reported on her death with an interview with her mother on October 31.

These deepfakes are devastating to families of victims as well as supporters of Israel and are now putting brands at risk.

advertisement

advertisement

TikTok has been having a difficult time moderating these deepfake videos of Hamas victims as brands struggle to steer clear of any controversial messages. The creators of the videos are anonymous.

The videos violate TikTok’s content-generation policies, which ban AI-generated videos of private citizens and minors. They are being seen by millions, according to Alex Barinka, Bloomberg News journalist.

In an interview, Barinka tells Bloomberg that an image of Louk spoke about the deceased woman’s mother and the attack, but in fact it was not Shani Louk. Barinka called it a “digital resurrection” in synthetic media, which brands could serve up nearby.

The deepfake video, Barinka said, racked up 7 million views.

Events such as this have likely prompted Facebook and Instagram parent company Meta along with IBM to join an industry AI alliance to share technology that could protect users and advertisers.

Meta and IBM are joining more than 40 companies to create an industry group dedicated to open-source AI work, aiming to share technology and reduce risks.

The coalition, AI Alliance, will focus on the responsible development of AI technology, including safety and security tools, according to a statement published Tuesday on IBM’s website.

The group also will look to increase the number of open-source AI models — rather than the proprietary systems favored by some companies — to develop new hardware and team up with academic researchers.

The 50 founding members include but are not limited to: AMD, Anyscale, CERN, Cerebras, Cleveland Clinic, Cornell University, Dartmouth, Dell Technologies, EPFL, ETH, Hugging Face, Imperial College London, Intel, INSAIT, Linux Foundation, MLCommons, NASA, NSF, Oracle, Partnership on AI, Red Hat, Roadzen, ServiceNow, Sony Group, and Stability AI.

All the alliances worldwide cannot prevent social-media outlets from posting inappropriate content. That will require moderation and enforcement from the company.

Next story loading loading..