Dutch GPT‑NL is ready for use: ‘The only language model that complies with the GDPR'
February 26, 2026
Members in the news: Tweakers – Two years ago,(in Dutch) the Ministry of Economic Affairs and Climate Policy allocated 13.5 million euros for the development of a Dutch AI language model, and now the time has come: GPT‑NL is ready for use.
The LLM, developed by research institute TNO in collaboration with SURF and the Netherlands Forensic Institute, is intended to offer universities, researchers, and public authorities a ‘sovereign’ alternative to existing AI tools.
TNO set itself the task of creating an AI bot that not only speaks Dutch but is also as privacy‑friendly as possible. For example, this meant that the model uses only responsibly sourced training data. That effort appears to have paid off: last month, GPT‑NL won a Privacy Award. The Dutch organizer of that award, Privacy First, described (in Dutch) GPT‑NL as the “first LLM worldwide that demonstrably complies with the requirements of the GDPR.”
Tweakers spoke with TNO about how GPT‑NL was developed and what it took to make the model as privacy‑friendly as possible.
Tweakers spoke with TNO about how GPT‑NL was developed and what it took to make the model as privacy‑friendly as possible.
On March 13, SURF is organizing a meeting for our members on “GPT‑NL: a homegrown AI model in practice.” The event has proven to be very popular, as registration is now completely full.
Read the full article here.
Share this post!