Not long ago, “deepfake” was a term reserved for viral novelty videos and dark corners of the internet. In 2026, it’s a marketing liability that has landed a Fortune 500 appliance brand and one of advertising’s most prestigious award shows in federal court.
As generative AI tools make it easier than ever to fabricate a convincing voice, face, or testimonial, the line between “creative use of AI” and “deceptive manipulation” has become a question every marketing team and every organization that hires one needs a clear answer to.
For nonprofits, schools, congregations, and public agencies whose entire value proposition rests on public trust, getting this wrong isn’t just a legal risk. It’s an existential one. The concept of ethical AI in marketing must guide our strategies moving forward.
What the Research Says About AI-Fabricated Content
A 2024 study from Cornell’s Johnson Graduate School of Management set out to measure just how convincingly generative AI could fabricate the kinds of content marketers rely on every day, such as reviews, testimonials, and other user-generated content.
The findings were sobering: the research demonstrated AI’s proficiency at fabricating disinformative user-generated content that closely mimics authentic content, and quantified how even minimal levels of this disinformation can disrupt marketing analytics and research. Perhaps most notably, the study found that standard detection techniques are insufficient to filter AI-generated disinformation from the data marketers rely on. The tools most agencies already use to gauge sentiment, monitor reviews, or benchmark competitors may be quietly compromised without anyone noticing.
The researchers’ recommendation wasn’t a single fix, but a layered one: better detection algorithms, renewed human oversight, and updated ethical and regulatory frameworks working together, because no single safeguard is sufficient on its own.
A Cautionary Tale From Cannes
If the academic research feels abstract, a recent real-world case makes the stakes concrete. In 2025, an advertising agency working for Whirlpool in Brazil used AI to alter footage from a 2018 TED Talk given by DeAndrea Salvador, a North Carolina state senator, changing her voice and the content of her remarks to promote appliances in São Paulo rather than the U.S. energy-affordability message she’d actually delivered.
The manipulated ad didn’t just run quietly; it won a Grand Prix award at the Cannes Lions International Festival of Creativity, one of advertising’s most prestigious honors, before anyone flagged that the “testimonial” was fabricated. According to Axios’s reporting, the ad used copyrighted footage of Salvador discussing energy affordability in the U.S. and added a line, seemingly in her own voice, specifically about São Paulo data, all without her knowledge or consent.
Once discovered, the award was withdrawn, the agency’s chief creative officer was dismissed, and Salvador filed suit, describing the situation as an “egregious lack of corporate ethics.” The case illustrates a hard truth: sophisticated organizations with legal and creative review processes still got this wrong, and the fallout touched a public official, a household-name brand, and one of the industry’s most respected award bodies all at once.
Why This Risk Is Only Growing
The Whirlpool case isn’t an isolated incident. Deepfake-enabled fraud, ranging from cloned executive voices authorizing fraudulent wire transfers to fabricated celebrity endorsements circulating on social media, has become common enough that cybersecurity firms now track it as a standard threat category rather than an emerging one.
The barrier to entry keeps dropping: a few seconds of publicly available audio or video, easily harvested from a webinar, a podcast, or a social post, is often enough to build a convincing synthetic likeness.
For marketing teams, that means the risk isn’t limited to intentionally creating deceptive content. It also includes the possibility that your organization’s leaders, donors, or spokespeople could themselves be impersonated by bad actors, or that a well-meaning creative shortcut, such as repurposing a public figure’s old interview footage, crosses an ethical and legal line without anyone fully realizing it until it’s too late.
Building Guardrails: What Ethical AI Marketing Looks Like
The organizations navigating this well share a few common practices. They obtain explicit, documented consent before using anyone’s real likeness, voice, or words in AI-assisted content, no matter how minor the edit may seem.
They maintain a human review step focused specifically on authenticity and consent, separate from the standard creative and legal review, because the Whirlpool case shows that those existing processes can still miss synthetic manipulation.
They disclose AI involvement clearly rather than letting audiences discover it after the fact. And they train staff to recognize the early warning signs of impersonation, both as content creators and as potential targets. We touched on several of these same disclosure principles in our earlier post on generative AI in visual content creation.
The throughline across every format, whether it’s an image, a voice, or a full video testimonial, is the same: transparency protects trust, and trust is the one asset no organization can afford to gamble with.
Why This Matters Most for Mission-Driven Organizations
For a nonprofit, congregation, school, or government agency, credibility is the foundation of relationships with donors, congregants, families, and constituents. Younger audiences in particular have developed what we’ve described elsewhere as an “authenticity radar” — a heightened sensitivity to anything that feels manufactured or manipulated, and a swift, public willingness to call it out. An organization that gets caught cutting corners with AI-generated testimonials or fabricated spokespeople risks the very trust that took years, sometimes generations, to build.
The good news is that ethical AI use and effective AI use are not in tension. As we’ve written about using AI and human judgment together in modern marketing, the strongest results come from treating AI as a tool that supports real people and real stories, not one that replaces or fabricates them.
If your organization wants a second opinion on how AI tools are being used in your marketing, or you’re building internal guidelines for responsible use, Dogwood Media Solutions can help you put practical guardrails in place before a shortcut becomes a liability.
