There's one question that keeps coming up across marketing teams: I've just created this image with AI—do I have to disclose it? The short answer is: it depends on what's in the image, not on the tool you used.
The AI Act does not require you to label everything created with AI. Starting on 2 August 2026, it requires disclosure only when AI-generated or AI-manipulated content could be mistaken for something real—such as people, scenes, or places that look like authentic photographs or videos. An abstract background or a product on a plain white background? No disclosure required. A model who doesn't actually exist? Yes.
Let's look at the three most common scenarios.
- On your ecommerce website. If you've simply added a colored background to a product photo, you're good to go. But if you use virtual models that look like real people, or you've altered the face or body of a real model using AI, you need to provide a clear disclosure. It should be visible next to the image or on the product page—not hidden away in the fine print at the bottom.
- On social media. The same rule applies, with one advantage: platforms like Instagram and TikTok already offer their own AI-generated content labels. Turn them on when posting AI-generated videos or deepfakes. It takes just one click and helps you stay compliant.
- In a banner ad. If it only contains typography, a product, or an illustration, no disclosure is needed. But if it features an AI-generated person who looks real, the disclosure must be clearly visible.
The real challenge isn't adding a label—it's knowing which images, among the hundreds you publish, actually require one and making sure every relevant image is properly disclosed.
Lawwwing has automated this process with its AI Sentinel Plan, which scans your website, detects AI-generated content, and automatically labels it through a floating widget so your users can clearly see which content was created with AI—without you having to do anything.