What is AI-generated in the App

English meanings for canonical names not in our curated dictionary at launch — roughly 14 500 of ~14 800 entries — were produced by AI. Translations of every English meaning into the 10 non-English locales we support (Dutch, German, French, Spanish, Italian, Portuguese, Simplified Chinese, Japanese, Korean, Modern Standard Arabic) were also AI-generated.

In total, roughly 97% of the meanings shown in Babinom are produced or translated by an AI model. The remaining ~3% are either editorially curated by us or sourced from public-domain references such as Wikidata.

Each meaning row in our database is tagged with its source — curated-core, external:wikidata, or openai:<model-name> — so we always know which content came from where.

How accurate is AI-generated content

AI language models produce plausible-sounding text. They are not authoritative reference works.

Specifically, meanings may be incomplete, oversimplified, or one-sided. Origins may be assigned to the most common tradition when multiple cultures use the same name. Translations may be grammatically correct but culturally unidiomatic. Religious, mythological, or historical references may be summarised in ways that experts in those traditions would refine or contest. Rare or recently-coined names may produce confident-sounding but invented etymologies.

We mark each meaning with the model's self-assessed confidence; meanings below a threshold are flagged in our database for human review.

What you should NOT use it for

Please do not rely on the meanings shown in Babinom for legal naming, birth-certificate, or registry decisions; religious naming traditions where specific scriptural or community guidance matters; genealogical research where source provenance is required; academic linguistic or historical work; or tattoos, family heirlooms, and other permanent commitments where accuracy matters.

For any of the above, please consult an authoritative reference such as a published etymological dictionary, religious authority, or qualified naming specialist.

Low-temperature generation

We use temperature 0.2 when generating meanings, which favours consistency over creativity.

Confidence flagging

The model self-rates each meaning. Low-confidence entries are flagged in our database for human review.

Output validation

We validate every output against basic format rules (length, no refusal phrases, no tautologies) before applying.

Source priority

Where both exist for the same name, human-curated and Wikidata-sourced meanings take precedence over AI-generated ones.

How to report an error

If you find a meaning that's wrong, misleading, or culturally insensitive, please email us at support@babinom.com with the name, the meaning you saw, and what's wrong with it. We review reports and update the catalog when justified.

Our use of AI providers

We currently use OpenAI as our language-model provider. Meanings and translations are generated server-side by us before reaching your device; we do not call OpenAI from your phone or share your personal data with them during generation. The names sent for enrichment are names already in our catalog, not data tied to your account.

For information on what data is collected and processed in connection with your use of Babinom, see our Privacy Policy.

Changes to this Disclaimer

We may update this Disclaimer as our use of AI evolves. Material changes will be noted in the App.

Contact

Questions about this Disclaimer: support@babinom.com.

Related documents: Terms of Service · Privacy Policy · Cookie Policy.

Babinom · Daan Vos · Cullera, Spain