Resources
FAQ & guides
Clear answers on pseudonymisation, using AI with sensitive data and how NYMETRA works.
Irreversible on one side, re-attributable on the other: what changes before processing a document with AI.
Read →Guide · AI & confidentialityGenerative AI and sensitive data: what are the risks and precautions?Limit what is sent, check the provider, keep human control.
Read →Guide · AI & confidentialityHow do you pseudonymise a document before sending it to ChatGPT, Claude or Gemini?Consistent aliases, a mapping kept separately.
Read →Guide · Lawyers & AILawyers and AI: professional secrecy, confidentiality and good practiceDefine the task, reduce the content, validate the result.
Read →Is pseudonymised data still personal data?
Yes, as long as re-identification remains possible with additional information. Re-attribution information must be kept separately and protected.
Is removing first and last names enough to anonymise a document?
No. An address, a rare job title, a date or a combination of items can be enough to identify someone.
What is a mapping table and why protect it?
It links aliases to the original information. In NYMETRA’s normal operation, it stays with you and is not sent to the AI provider.
What do GPT, Claude or Gemini receive?
The pseudonymised version and what the request needs. Never the mapping table.
Does NYMETRA work like a subscription?
No: a one-off purchase, a right to use the version acquired, and 12 months of Basic included. Renewing the service is optional.
On how many devices can I use NYMETRA?
Each user can activate NYMETRA on two work devices. A device can be deactivated to activate another.
No results.Ask us →