For gamers, it’s all everybody’s talking about: The leaked audio of U.S. president Joe Biden and former President Trump in a game of Overwatch on voice chat.
It’s not real, of course, it's a deepfake. But in the battle of brand safety issues, deepfaked audio is the next frontier.
Adweek reports today that voice-based memes fueled by content created through generative AI are the latest hits on social media platforms, especially TikTok and YouTube.
Most of these AI-generated videos portray politicians, like Donald Trump and Joe Biden, playing video games — but many videos have their voices spreading misinformation and conspiracy theories, according to brand safety ad-tech firm Zefr's research.
A recent video example shows the pair trash-talking while gaming, in which AI-generated Biden asks for help selling his cocaine business, and Trump claims “America is tyranny and fascism.” And your brand's ads may be running alongside this content.
🚀 The Rise of Voiced-Based Memes
Data indicates that so far this year AI-generated videos portraying politicians across platforms have already exceeded the total volume of 2022 videos by over 130%.
Zefr reports that these videos have received over 11 million views and over 2 million likes, and over 90% of them are currently monetizable, meaning advertisers are running ads adjacent to them.
The report also found that misinformation videos were scripted using ChatGPT.
The company couldn't disclose how much money was being spent on ads that appeared next to AI-generated audio misinformation.
Adweek notes that advertisers can use keyword blocklists or AI-driven contextual tools from third-party verification companies like DoubleVerify and Integral Ad Science to filter ads that are brand-safe and brand-suitable outside of TikTok and Youtube, where these videos are currently on the rise.
Even so, a spokesperson for Zefr said that tools, including keyword blocklists and static inclusion lists, can be ineffective on social platforms that require content analysis beyond web site domains.
👀 Sexual Deepfakes of Celebrities Run Rampant in Ads
Meanwhile, sexual deepfakes Emma Watson and Scarlett Johansson ran rampant across Facebook and Instagram ads this week as part of a campaign for an AI deepfake app.
NBC News reported that 230 videos were run across Meta's platforms. According to the report, the campaign was an advertisement for the app FaceMega, which lets users swap out faces in videos with pictures uploaded to the app for $8 a week.
Images: YouTube / Screenshot: laurenbarton03/Gizmodo, Frazer Harrison
