AI tech results in explosion of deepfake videos online | DW News

AI Tech Results in Explosion of Deepfake Videos Online | DW News

With AI technology developing at breakneck speed, experts are warning of an explosion of deepfake online videos. Many people are falling victim, with their identities being used without consent to create digital content, sometimes with dubious motives.

Olga Loek's Experience

We meet a woman, Olga Loek, whose face has been used to make thousands of fake videos. About a month after the 21-year-old Ukrainian started her YouTube channel, she began receiving strange messages asking if she spoke Mandarin. Initially, she thought it was a joke, but then decided to watch some of the videos people were sharing with her.

"I saw that this is literally like my face speaking Mandarin, and in the background, I'm seeing Kremlin and Moscow, and I'm talking about things like how great Russia and China are."

Olga's image is being used to spread Russian propaganda and sell Russian products in China. She translates one of the videos into English:

"I really appreciate China while everyone else is moving away from Russia. Only China is behind with quiet support. When other countries reject Russian stuff, it's your support that helps Russia meet its immediate needs."

The Scale of the Problem

Olga has multiple digital doppelgangers, none of whom exist. One appears to be a Russian woman named Natasha who wants to support Russia's war by selling products from her homeland. One company in China told Olga that some 5,000 videos had been created with her face on their platform alone.

Ease of Deepfake Creation in China

xmov, a company in China that develops advanced AI technology, noted the tech behind such deepfake videos is very common in China. The company was not involved in Olga's situation, but the CEO Jim Chai says the videos show just how easy deepfakes are to produce.

"The only AI technology used here is lip-syncing. If ordinary people don't pay attention to the lips, they may not notice the difference."

Expert Insights

Henry Ajder, an expert on AI and deepfakes who advises institutions like the European Commission as well as tech companies like Meta and Adobe, shared his perspective on the issue.

"These cases are still quite rare, especially the ones where someone is turned into a digital puppet for promoting services. But the tools only require a few minutes of voice audio and video, so the victim profile could be anyone."

Preventive Measures

With AI tools requiring minimal data and achieving high sophistication, it's challenging to prevent such instances. Ajder suggests that the onus should be on companies developing these tools to introduce better safety measures, like biometric authentication.

Global Regulatory Landscape

The global regulatory picture is irregular, with few countries passing formal legislation on AI technologies. The EU has taken a comprehensive approach, but enforcement is still a challenge in a digital world that doesn’t respect national boundaries.

Positive Uses

Though deepfake technology presents risks, it also offers positive benefits. For instance, voice cloning tools can help people who’ve lost their ability to speak or create engaging training materials using avatars. The same baseline generative technologies used to create deepfakes can also have pro-social applications.

Henry Ajder concludes:

"There are a lot of things with synthetic media and AI-generated content that we can be excited about. However, like most dual-use technologies, there are potential harms."

Keywords

• AI technology
• deepfake videos
• digital content
• Russian propaganda
• synthetic media
• voice cloning tools
• generative AI
• biometric authentication
• EU regulation

FAQs

Q: What triggered Olga Loek to investigate deepfake videos using her face?
A: Olga began receiving strange messages asking if she spoke Mandarin, prompting her to investigate.

Q: What kind of content were Olga's deepfake videos promoting?
A: They were promoting Russian propaganda and selling Russian products in China.

Q: How prevalent is deepfake technology in China?
A: According to xmov, deepfake technology and its creation tools are very common and easy to use in China.

Q: What measures can individuals take to prevent their likeness from being misused in deepfakes?
A: It's challenging for individuals to prevent this misuse; instead, companies should introduce safety measures like biometric authentication.

Q: What is the EU's stance on regulating AI and deepfakes?
A: The EU has taken a comprehensive and rigorous approach but faces challenges in consistent enforcement.

Q: Are there any positive uses for deepfake technology?
A: Yes, deepfake technology can be used for beneficial purposes such as helping individuals who’ve lost their ability to speak and creating engaging training materials.

By reconstructing the given script into an article format and providing summaries and FAQs, the information becomes easier to digest while addressing key aspects of the deepfake issue.