There is a date few marketing departments have marked in their calendars, yet it should concern them: 2 August 2026. That day, one part of the European regulation on artificial intelligence, the AI Act, comes into effect. Its Article 50 imposes a transparency obligation: certain content generated or manipulated by AI will now have to be flagged as such.
For two years now, we have been told that synthetic imagery will absorb everything: the shoots, the actors, the sets, the craft. From Cannes, we read this deadline differently. It does not signal the end of filmed images. It is a reminder of their value.
What the law says, and what it doesn’t
Let’s be fair, because the subject lends itself to anxious shortcuts. The AI Act does not require labelling every piece of content made with AI. Article 50 mainly targets deepfakes, those images, sounds and videos that imitate real people or real situations, along with content that touches on information and the public interest.
To support its rollout, the European Commission published on 10 June 2026 a code of good practice on the marking of AI-generated content. A tool, voluntary for now, that helps players translate the spirit of the text into concrete actions: marking, metadata, visible disclosures.
In short: no one will stamp “generated by AI” on every post. But as soon as a piece of content can mislead about what is real, flagging it becomes the norm. And that is exactly where the calculation shifts for brands.
For a brand, synthetic becomes a liability
As long as the fake stayed invisible, it was tempting. A scene impossible to shoot, an unreachable location, an impossible deadline: AI promised to fill the gap without anyone knowing. The moment you have to admit it, the equation flips.
Content flagged as artificial sends a message few brands want to carry: what you are seeing did not happen. In a climate where trust is already fragile, that is a credibility risk as much as a legal one. Audiences quickly learn to spot the small disclosure, and to wonder what it covers.
A filmed image, on the other hand, has nothing to declare. It needs no asterisk. The real carries no legal notice.
And the trend will only intensify. The European framework is built to tighten, not to loosen. Betting today on genuinely filmed images means building a library of content that will never become a compliance problem: nothing to reclassify, nothing to flag, nothing to pull the day the requirements step up a level.
The real becomes a competitive advantage
Let’s say it plainly: defending the shoot today is not purist nostalgia. It is a cold reading of the market.
When anyone can manufacture synthetic “beauty” in seconds, what becomes rare, and therefore valuable, is proof. Real light on a real face. Emotion captured, not simulated. The product, the place, the gesture, filmed where they actually exist. A camera facing the real says something no prompt reproduces: it happened, we were there.
That has been our ground since day one. When we shoot a brand film on the French Riviera, cover an event in Cannes or stage a product, we are not trying to imitate the real. We film it. That gap, long filed away as an aesthetic preference, is becoming an argument for trust.
Filming the real, more than a principle
AI has its place in our craft: speeding up what should be sped up, lightening the technical load, freeing time for the eye. But when it comes to what a brand shows the world and signs with its name, the transparency imposed by law reshuffles the deck in favour of those who have nothing to hide.
On August 2nd, brands will have to say when something is fake. We would rather spend that time filming what’s real.
Beauty, now, everyone promises it. The real, you still have to go and get it, camera on the shoulder. Pas que du beau. Du vrai. Du BOW.
A real image to shoot, a brand to embody for good? Let’s talk.