The “Black Box” era of marketing is over. As of early 2025, with the EU AI Act fully enforceable and similar regulations pending in California and New York, transparency is no longer just a nice-to-have—it’s a legal baseline.

But legal compliance is the floor, not the ceiling. The real question for brands isn’t “Do we have to disclose this?”, but “What happens to our trust if we don’t?”

The Trust Gap

We are living in a low-trust media environment. Consumers assume everything is filtered, edited, or faked until proven otherwise. When you use Midjourney to generate a product backdrop or ChatGPT to write a newsletter, and you don’t declare it, you aren’t just saving time—you are withdrawing from your brand’s trust bank.

The 2024 Edelman Trust Barometer found that 73% of consumers expect brands to clearly disclose AI usage. More critically, 68% said they would stop buying from a brand that “deceived them” about AI-generated content. The fallout from Coca-Cola’s AI-generated holiday ad in December 2024—which faced massive backlash despite the company’s transparent disclosure—shows that even honest AI use can be polarizing. But the brands that got burned worse? The ones who tried to hide it.

Toys “R” Us, by contrast, turned their AI-generated brand origin video into a transparency win by releasing a behind-the-scenes documentary about the Sora-powered production process. The campaign earned 12 million impressions and a Cannes Lions shortlist nomination—not despite the AI disclosure, but because of it.

The “Verified Human” Badge

Interestingly, we are seeing the emergence of a “Human-Verified” transparency movement. Brands are beginning to use “No AI” watermarks as a premium differentiator—a digital equivalent of “Handcrafted” or “Organic.”

Conversely, the “AI-Assisted” tag is losing its stigma. It is becoming a sign of sophistication. “This image was visualized with AI” tells the customer: We are innovative, but we are honest.

The smartest brands are practicing what I call The 3-Second Disclosure Test: If a consumer can’t identify that content is AI-generated within 3 seconds of encountering it—through a clear label, watermark, or caption tag—you’ve failed transparency.

Best Practices for 2025

So, where is the line?

1. High-Risk vs. Low-Risk: If the AI is essentially acting as a spellchecker or a brainstorming partner, disclosure is optional. If the AI is generating the core value (the artwork, the voice, the face), disclosure is mandatory. EU regulations specifically target “high-risk” AI in content creation, customer interaction, and decision-making.

2. Labeling: Don’t hide the disclosure in the terms of service. Put it in the caption. Use platform-native tags: Instagram and TikTok now support #AIGenerated labels; LinkedIn has an “AI-Assisted” post tag. Meta’s “Made with AI” label became mandatory in February 2025.

3. The “Sandwich” Method: Frame AI content with human context. “I used AI to visualize this concept [AI Image], but here is why this concept matters to me [Human Story].” Human framing + AI execution + Human interpretation = Transparent value.

The Bottom Line

Transparency doesn’t kill the magic. It builds the trust required for the magic to work. In an age where every brand has access to the same AI tools, your competitive advantage isn’t the technology—it’s the integrity.

The brands that win in 2025 won’t be the ones that hide their AI usage. They’ll be the ones who are so confident in their human judgment that they can afford to show you exactly how the AI sausage is made.

Because trust isn’t just good ethics. It’s brand equity.