Analysis · 5 October 2026

AI-generated naming: 9 out of 12 names already registered or at risk

Original title, in French: “Naming par IA : 9 noms sur 12 déjà déposés ou à risque”

Publication: Journal du Net (France)
Author: Florian Bourgeois
Published on: 5 October 2026 · read in the page structured data

Summary

A consultant tested four AI assistants (ChatGPT, Gemini, Mistral, Claude) on a real naming brief for an online shop selling premium cat accessories, obtaining 160 name suggestions. Of 12 names checked in the TMview trade mark database, 9 are already registered or present a risk of conflict. Despite an instruction to avoid clichés and a follow-up prompt, the tools keep producing names built on expected vocabulary (fel-, cat, chat, paw…) or on similar suffixes (-ora, -elle). The author points out that no tool makes a genuine strategic choice and that ChatGPT wrongly claimed to have checked a trade mark's availability. The conclusion recommends using AI to explore options, not to choose a trade mark name or check it legally.

Summary based on the full text of the article.

Read the article on Journal du Net →

Figures cited

9 names out of 12

Tested names already registered or presenting a risk of similarity with an existing trade mark

12 names from 4 AI assistants, checked in TMview (France/EU), October 2026

« Résultat du test : 9 noms sur 12 déjà déposés ou à risque »
160 names

Total number of names generated by the four AI assistants tested

4 AI assistants (ChatGPT, Gemini, Mistral, Claude), 20 names each, with follow-up prompt

« J'ai soumis le brief réel d'une boutique en ligne à quatre assistants IA, soit 160 noms au total. »
67 names out of 80

First-round names containing the expected cat-related vocabulary despite the anti-cliché instruction

80 first-round names (4 assistants x 20 names)

« Pourtant, 67 noms sur 80 au premier tour contiennent le vocabulaire attendu du chat : fel-, cat, chat, purr, paw, mew, neko, minou. »
45 names out of 80

Names containing the expected cat-related vocabulary after the anti-cliché follow-up

80 second-round names, after follow-up

« La proportion tombe à 45 sur 80, mais pas partout »
19 names out of 20

Gemini names built on cat vocabulary after the follow-up (increase despite the announced correction)

Gemini list, second round

« Gemini, après avoir admis ses répétitions, passe de 17 à 19 noms sur 20 construits sur le lexique du chat. »
41 names out of 80

Second-round names ending in a suffix such as -a, -ora, -ia or -elle

80 second-round names

« 41 noms du second tour se terminent en -a, -ora, -ia ou -elle (Luxora, Veloria, Norelle, Velissia). »

Classification

Country: France
Company size: All sizes
Industries: Consulting, legal, accounting, business services · Media, communications, marketing
Companies: Anthropic (Claude) (supplier), OpenAI (ChatGPT) (supplier), Google (Gemini) (supplier), Mistral AI (supplier)
Keywords: naming, name generation, industrial property, trade mark registration, chatbots, branding

Entry written by Marsanne Studio from the cited article, with the help of an AI model (claude-sonnet-5). The rights to the article and its illustrations belong to Journal du Net and its author. Entry translated from French by Claude Sonnet 5.5.