Image created using DALL·E 3 with this prompt: Create an image of a future that is overwhelmed by advertising. Everywhere you look, you see advertisements being endorsed by celebrities and influencers. It’s hard to tell what’s real and what’s fake. It’s overwhelming and overstimulating. Create a sense of claustrophobia. Neon colors and hyper saturated. Filled with digital screens and other futuristic technology. Aspect ratio: 16×9. [then prompted five more times: “Great. Make it even more claustrophobic.”]
Deepfakes are all fun and games until a synthetic Taylor Swift starts promising free cookware on TikTok – then it requires Federal legislation.
This problem is rampant. YouTube has taken action against a significant number of AI-generated scam advertisements featuring unwitting celebrities by deleting more than 1,000 videos following a 404 Media investigation. These videos, which amassed nearly 200 million views, employed AI to fabricate endorsements from a whole slew of high-profile individuals (the Medicare videos with fake Steve Harvey and fake Joe Rogan were particularly good). Despite these removals, the platform continues to host similar deceptive content, including an ad using an AI-generated voice of Donald Trump to disseminate misinformation and another featuring a fabricated endorsement from Jennifer Aniston for a fraudulent offer.
This issue gained attention after an undergraduate student compiled and shared a list of such videos with 404 Media, prompting YouTube to not only remove the identified videos but also to ban the responsible accounts. The student’s anonymity was preserved due to concerns over potential impacts on job prospects.
In response to the proliferation of these deepfake ads, YouTube (and by extension Google) acknowledged the challenge and is actively working to enhance its detection and enforcement mechanisms. A Google spokesperson said the company is investing in advanced image recognition technology and improving models to identify unauthorized celebrity depictions in videos. This effort is part of a broader initiative by Google, which reported the removal of over 142 million ads in 2022 for policy violations related to misrepresentation.
The issue is similar to another one currently plaguing X, where trolls have flooded the platform with graphic Taylor Swift AI fakes. In both cases, it’s a game of Whac-A-Mole where the scammers only need to fool you once. I would ask you to re-read my essay Default to Distrust: A New Paradigm for Engaging with Generative AI Models. It’s our only defense.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.