When AI Ads Steal Your Face: The Deepfake Advertising Crisis

Imagine waking up to discover your likeness plastered across an advertisement—except you never agreed to it, never signed a contract, and worst of all, the version of you being sold to millions of people doesn't actually look like you anymore. This isn't a dystopian nightmare. It's happening right now, and YouTuber Elliot Choy became the unwitting face of this emerging crisis when his video was stolen, AI-altered to change his race, and repurposed in a Kalshi cryptocurrency advertisement without his knowledge or consent.

The incident involving Choy represents a critical inflection point in the advertising industry. What was once confined to science fiction—the ability to convincingly manipulate someone's appearance and likeness at scale—is now cheap, accessible, and increasingly weaponized by companies seeking to reduce production costs and maximize reach. The Kalshi case wasn't an isolated mistake or a rogue actor; instead, it exposed a troubling pattern of AI-generated advertising content being deployed without creator permission, raising urgent questions about copyright infringement, identity theft, and the future of authentic representation in commercial media.

The Choy Case: When Your Face Becomes a Corporate Asset

Elliot Choy, a content creator with a substantial following, discovered that a video he had created was being used in a Kalshi advertisement promoting their prediction market platform. The shock deepened when he realized the video wasn't simply repurposed—it had been algorithmically altered to change his race, transforming his appearance into something entirely different while maintaining the essence of his original content. The advertisement ran across multiple platforms, potentially reaching millions of viewers, all without Choy's permission or knowledge.

When Choy made the discovery public, Kalshi initially claimed they had purchased the rights to use the content, a statement that immediately raised red flags about where they actually obtained the material and whether proper vetting occurred. The company's explanation—that they believed they had legitimate licensing—doesn't erase the fundamental violation: an AI tool was used to alter someone's identity, and that altered version was monetized without consent. The incident sparked significant discussion in creator communities, forcing uncomfortable conversations about how easily AI tools can be weaponized against content creators.

What makes this case particularly egregious is the identity manipulation component. It's one thing to reuse someone's likeness without permission—that's copyright infringement, clearly problematic. It's another thing entirely to algorithmically alter that person's race before using their image in advertising. This crosses from intellectual property violation into something more sinister: it suggests the person being depicted in the ad is someone they're not, which introduces questions about fraudulent representation and potential discrimination.

The Broader Pattern: Why Kalshi Might Just Be the Beginning

Industry observers and digital rights advocates argue that the Kalshi-Choy incident likely represents the tip of a much larger iceberg. Companies have financial incentives to use AI-generated or AI-manipulated advertising content: they eliminate expensive production costs, avoid negotiating licensing fees with creators, sidestep union regulations, and can rapidly test different demographic variations of the same ad. When the technology makes all of this possible with minimal detection risk, the temptation for cost-cutting marketing departments becomes enormous.

Kalshi's claim that they purchased rights to the content raises critical questions about the supply chains for AI-generated advertising material. Who are these suppliers? Are they actually acquiring legitimate rights, or are they scraping content from the internet and selling it to the highest bidder? A recent FTC investigation into advertising practices suggests that companies often operate with minimal oversight regarding content sourcing, particularly when that content is algorithmically generated or modified.

The economics of this problem create a perverse incentive structure. As AI image and video generation tools become more sophisticated and cheaper to operate, the cost of creating "authentic-looking" advertising content approaches zero. A marketing team could theoretically generate dozens of ad variants—each depicting different demographics, ages, or ethnicities—all from a single stolen source video. The original creator receives nothing, while the company benefits from massive cost savings and increased conversion rates through better demographic targeting.

The Regulatory Vacuum: Where's the Protection?

Perhaps the most alarming aspect of the deepfake advertising crisis is the almost complete absence of regulatory frameworks to address it. Unlike traditional copyright law, which has existed for centuries, or even the Digital Millennium Copyright Act (DMCA), which was written in 1998 to address digital content protection, there are no clear legal guidelines specifically addressing AI-generated advertisements or AI-manipulated content used without consent.

The Federal Trade Commission (FTC) has begun issuing guidance about AI and advertising, but these recommendations remain largely toothless without legislation backing them up. The FTC can issue fines for deceptive practices, but what exactly is deceptive about an AI-generated ad? If the company didn't explicitly claim the person in the ad was real, have they technically lied? These legal gray areas persist because lawmakers have struggled to keep pace with technological change. FTC guidance on AI advertising claims exists, but it focuses more on preventing exaggeration about product benefits than on protecting individuals from unauthorized use of their likeness.

The European Union's AI Act represents one of the few legislative attempts to create comprehensive rules governing AI systems, but even that sweeping regulation doesn't fully address the specific issue of AI-generated advertising without consent. Individual states in the United States have begun passing laws protecting people from deepfake pornography and non-consensual intimate images, but advertising deepfakes exist in a different category entirely—and currently occupy a legal blind spot.

Creators and individuals harmed by unauthorized AI-generated advertising content must currently rely on traditional copyright law, right of publicity statutes, or terms of service violations—tools designed for an era before AI could convincingly impersonate anyone. A creator might successfully prove copyright infringement if the source material was stolen, but what recourse do they have if someone uses their face without their likeness being copyrighted? What happens when the altered version of them—the race-changed version—becomes the publicly associated image?

The absence of regulation has created a Wild West environment where companies can experiment with AI advertising largely consequence-free. Penalties, when they occur, are typically treated as minor business costs rather than serious deterrents. This cost-benefit analysis heavily favors companies willing to take the risk: if there's a 10 percent chance of getting caught and facing a fine that amounts to a small fraction of the money saved through not paying creators, the math still works in their favor.

Moving forward, stakeholders across the industry recognize that action is necessary. Creators need tools to detect and prove unauthorized use of their likeness. Platforms need clearer policies about what types of AI-generated content are acceptable. Regulators need to establish concrete rules distinguishing between legitimate AI-generated advertising and deepfake content created without consent. The Elliot Choy case serves as a warning that the current system is inadequate—and that without swift intervention, unauthorized AI-generated advertising will only become more prevalent as the technology improves and costs continue to plummet.