The age of image manipulation is not new. But what has changed—what demands our attention now—is how easily we can manufacture truth. Welcome to the era of deepfakes.
Deepfakes are videos or images created using AI that show people doing or saying things they never actually did. By August 2023, tools like DeepFaceLab and D-ID had made it possible to create realistic face swaps with nothing more than a laptop and a few photos.
In entertainment or parody, this might seem harmless, but in design—particularly when trust is part of the user experience—the stakes rise sharply.
Let’s say you’re building an ad campaign. You use AI to generate an image of a celebrity holding your product. It looks amazing, but it’s not real. Is that ethical? Legal? Could you even tell it’s fake a year from now?
According to a 2023 DeepMind report, over 60 percent of people cannot distinguish between a real and a synthetic image when shown side-by-side for less than five seconds. That’s a design problem. And an ethical one.
Three Ethical Guardrails for Deepfake-Aware Design
- Transparency First
If you used AI to generate faces or voices, disclose it. A small badge or text label builds trust—and may soon be required by law (EU AI Act). - Consent Matters
Never use a person’s likeness, voice, or brand—real or synthetic—without permission. This includes composite or “lookalike” imagery. - Context Is Everything
A parody video? Fine. A fake testimonial from a “customer?” Unacceptable. Design is persuasion and with that comes responsibility.
The question isn’t can we make deepfakes; it’s should we? And if we do, how do we keep design honest in a world where visuals lie?
