Shaping.
Proving authenticity rather than styling it
Étude7 min

Proving authenticity rather than styling it

Facing counterfeiting and European regulatory ambiguity over AI-generated recommendations, houses such as Richemont and LVMH are reorienting artificial intelligence toward digital identity and the provenance of pieces. Trust is now built through technological proof of origin rather than the promise of style.

In Geneva as in Paris a movement is taking shape, quiet but clear. Several major European houses are reorienting their use of artificial intelligence. Where the effort once focused on style, on content generation, on personalized suggestion, it now shifts to a less visible and more decisive ground: the digital identity of pieces and the proof of their provenance.

The context explains this refocus. Counterfeiting has industrialized. Copies now carry signatures close to the original, convincing packaging, certificates that resemble those issued by the houses. The distinction between the real and the false no longer holds to the naked eye. It holds to data one can follow, verify, authenticate. That is precisely the ground where artificial intelligence becomes useful, not to imitate but to trace.

Added to this is a European regulatory ambiguity. The AI regulation requires, in certain cases, a disclosure when recommendations are generated by a system. The line between legitimate suggestion and subtle manipulation remains poorly fixed. Rather than wait for the framework to clarify, some houses have chosen to withdraw their AI stylist features outright from their European apps. The withdrawal is a signal. It indicates that, for now, the risk of opacity outweighs the benefit of automated personalization.

What this withdrawal reveals goes beyond compliance caution. It marks a change of doctrine. The question is no longer what AI can produce but what it can guarantee. For years luxury explored AI through the angle of style: generating an atmosphere, composing a visual, drafting a line. That use assumed value lived in appearance. The current moment recalls that in luxury value lives first in origin. A visual can be copied. A provenance, once technologically sealed, cannot.

The European groups, Richemont in Geneva, LVMH in Paris, are thus investing in digital identity. Each piece is tied to a passport, an immutable record, a trace that links the object to its workshop, to its material, to its gesture. The client no longer buys only an object and a brand. They buy a chain of proofs. This chain turns authenticity from an assertion into a verification.

The consequence for the relationship of trust is deep. Luxury long rested on promise. Promise of quality, promise of origin, promise of rarity. Promise assumes the client's faith in the brand. Proof, by contrast, dispenses with faith. It offers a verifiable, enforceable, transferable assurance. It is a change of nature, not of degree. Where promise asks you to believe, proof lets you know.

This shift also redefines the role of marketing. The marketing function of luxury was built on the styling of authenticity: telling the workshop, staging the gesture, sublimating the material. That art remains, but it yields to a sterner demand. Saying is no longer enough. Showing, proving, tracing become the core of the discourse. The houses that understand this will turn their communication into demonstration. Those that refuse will keep styling an authenticity their competitors have already proven.

There is in this movement a form of rigor that suits European luxury. Luxury was never excess. It was exactness. Putting proof back at the center is returning to that founding demand. Artificial intelligence, diverted from styling toward authentication, recovers a noble function: not to imitate the beautiful but to guarantee the true. That is perhaps the most precious service it can render to a sector whose value rests, in the end, on trust.

The coming years will tell which houses made proof an argument. The market is already deciding. In luxury what can be proven sells. What is merely styled is doubted.

© 2026 Shaping Digital Generation