Second-Hand Luxury: Your Tools Are Approving Fake Authenticity Certificates

AI-generated fake certificates are already passing checks

A seller submits a listing for an Hermès Birkin bag at €8,500. The photo is perfect. The attached invoice shows a readable serial number, flawless typography, and the brand logo centered to the pixel. The curation team processes 25,000 items that day. They approve the listing in seconds.

What nobody sees: the photo was generated in four minutes by an AI image tool. The serial number on the invoice was automatically replaced in under a second, at a cost of less than one euro cent. The bag shipped to the buyer is an industrial replica. The fraud is completed before the listing even goes live.

This scenario is not hypothetical. It happens every day on major luxury resale platforms. The tools to detect it exist — provided you know where to look.

The second-hand luxury market is attracting a new generation of counterfeiters

The global luxury goods market fell by 2% in 2024 to €1,478 billion [Le Particulier — Luxe : un marché en recul en 2024 (2025)], with an estimated loss of 50 million consumers over two years [Le Particulier — Luxe : un marché en recul en 2024 (2025)].

Demand is shifting massively to second-hand. This market now represents €35 billion globally and is growing by +15.5% per year [Bpifrance Big Media — 5 chiffres sur le marché de la seconde main (2024)]. Second-hand fashion revenues could reach $350 billion by 2027, growing eleven times faster than new sales [Bpifrance Big Media — 5 chiffres sur le marché de la seconde main (2025)].

Watchmaking follows the same trajectory. The secondary market for prestige watches is projected at $35 billion by 2026 [AI CERTs News — Luxury Authentication AI transforms watch verification (2025)]. Platforms such as Chrono24 manage more than 600,000 watches in their catalog, with between 9 and 10 million unique visitors per month [McKinsey — The State of Fashion Watches & Jewellery (2024)].

This concentration of value on digital platforms has transformed the nature of risk. Counterfeiters no longer target only the physical object. They target the digital trust contract that represents it: the photo, the certificate, the invoice.

Globally, trade in counterfeit products represents 3.3% of world trade [Trust-Place — La contrefaçon investit le marché du luxe de seconde main (2024)]. In the European Union, counterfeiting costs the economy nearly €16 billion each year [FashionUnited — Louis Vuitton est la marque de luxe la plus contrefaite (2025)].

Why your current verification processes have a blind spot

A business model built on trust in the digital image

Luxury resale platforms built their value proposition around a simple assumption: the image submitted by the seller faithfully represents the real item. That assumption was reasonable when creating a fake image required hours of work and left visible traces.

That is no longer the case. AI image-generation tools — such as Midjourney or Stable Diffusion — now make it possible to produce synthetic photographs that visual experts cannot distinguish from a real photo [GIJN — Reporter's Guide to Detecting AI-Generated Content (2024)]. A counterfeiter can generate a coherent series of photos of an Hermès Birkin — multiple angles, realistic indoor setting, natural light and shadow — in under five minutes.

Volumes that make human review impossible

Vestiaire Collective processes up to 25,000 items submitted by sellers every day [Marks & Clerk — Vestiaire Collective's Trust Report 2022 (2022)]. Even with large teams, each human verifier has only a few seconds per listing. At that pace, spotting an AI-generated image or an invoice whose serial number was automatically replaced is humanly impossible.

In 2024, around 10% of items submitted on Vestiaire Collective were rejected during the initial digital verification phase [Vestiaire Collective — Trusted Pre-Loved Luxury Authentication Standards (2024)]. This rate measures detected rejections — not sophisticated fraud that slipped through.

Physical counterfeits reaching industrial-grade quality

The documentary threat rests on an increasingly solid physical base. Around 40 million fake watches are produced every year worldwide [Claims Journal — Calling Time on Luxury Watch Fraud: A Digital Counter-Revolution (2025)]. Among them, 10% are superfakesVery high-quality counterfeits made with the same materials as the originals — steel, sapphire crystal, movement — and impossible to identify with the naked eye without specialized precision tools.: replicas that only a master watchmaker equipped with a microscope can distinguish from the original [Claims Journal — Calling Time on Luxury Watch Fraud: A Digital Counter-Revolution (2025)].

In this context, the only thing separating a superfake from a legitimate sale on a digital platform is the quality of the accompanying document. And that is precisely the document that AI can now falsify in a fraction of a second.

Why existing safeguards are no longer enough

Market leaders have invested heavily in security. Vestiaire Collective deploys algorithmic moderation and physical authentication in warehouses. Since April 2024, Chrono24 has offered its "Chrono24 Certified" program, which includes opening the case, checking the movement, and consulting stolen-watch databases [Chrono24 — Chrono24 Debuts New Certified Program (2024)]. These approaches are rigorous.

But they share a common flaw: they evaluate what the image shows, not how it was built.

A visual recognition algorithm can confirm that the photo looks like a real Gucci bag. It cannot detect that the photo was generated by an AI tool.

A human expert can validate that the typography on a warranty card appears consistent. They cannot see that the serial number was automatically replaced — an operation that costs $0.01 and takes less than a second [arXiv — AIForge-Doc: A Benchmark for Detecting AI-Forged Tampering in Financial and Form Documents (2025)].

Visual moderation — human or classic algorithmic — is now blind to AI-generated documentary fraud. Not due to a lack of skill, but because it was not designed to analyze the mathematical reality of a digital file.

Direct peer-to-peer shipping worsens this vulnerability. In this flow, there is no physical authentication in a warehouse. Digital moderation of the initial photos becomes the only barrier. If it approves a synthetic image, the counterfeit reaches the buyer directly [Vestiaire Collective FAQ — Are counterfeit products frequently found on Vestiaire Collective? (2024)].

Legal and compliance framework: what matters most

The legal consequences of a forged document always depend on the facts, the sector involved, the applicable qualification, and the competent jurisdiction. In practice, the main issue for an organization is to be able to demonstrate a proportionate, traceable, and well-documented verification process, with human review whenever a decision may have a significant effect.

The controls described here should therefore be understood as risk-management, compliance, and evidence-preservation measures. Any final blocking decision, report, contractual sanction, or legal action should still be validated by the relevant legal or compliance teams.