Deepfake Dilemma: Celebrities Targeted by AI Scams
No attachments for this post
Tom Hanks recently sounded the alarm on Instagram about a deepfake video of him being misused to advertise a dental plan. "BEWARE!! An AI rendition of me is promoting some dental scheme. I’m not associated with it," Hanks alerted. Gayle King of CBS faced a similar situation when her likeness was used to push a weight-loss product. She promptly shared the manipulated video labeled "Fake Video" and clarified, "I’ve never endorsed or used this product! Don’t be misled by these AI manipulations."
In a similar vein, YouTube sensation MrBeast alerted his followers on social media platform X about a deceptive video of him, doctored to market a bogus iPhone 15 giveaway. He posed a pressing question, "Can social media platforms manage the surge of AI deepfakes? This is gravely concerning."
Though we haven’t viewed the Hanks video, examples from King and MrBeast suggest scammers adapted existing celebrity footage, adjusting lip motions to sync with AI-crafted voice imitations based on public vocal samples.
These events contribute to the ongoing discourse about AI's ethical and legal role in media and entertainment. The Writers Guild of America's recent strike highlighted AI's controversial nature. SAG-AFTRA, the Hollywood actors' union, has voiced concerns about unpaid and unauthorized digital actor reproductions. Zelda Williams also made headlines, criticizing unauthorized uses of her late father, Robin Williams', voice.
Convincing deepfakes present a looming challenge, potentially compromising trust and questioning the authenticity of communication means. Solutions are under exploration: Google and OpenAI intend to watermark AI-produced content and insert metadata for origin tracking. Yet, such measures have been historically bypassed, and there are watermark-free open-source AI tools.
Regulating AI software might deprive legitimate researchers of generative AI tools, leaving malicious actors undeterred. To combat this, social media platforms must intensify moderation efforts, promptly addressing user-flagged dubious content.
Comments on this post
No comments have been added for this post.
You must be logged in to make a comment.