Universiteit Leiden

nl en

Research project

Mapping Out Multilingual Use of ChatGPT in the Netherlands

This project investigates how multilingual people in the Netherlands use ChatGPT: which languages they choose when they prompt, when they switch between languages, and which beliefs about language guide those choices.

Duration
2026 - 2027
Contact
Naomi Truan
Funding
NWO, Open Competition SSH XS
Photo by Priscilla Du Preez 🇨🇦 on Unsplash
Photo by Priscilla Du Preez 🇨🇦 on Unsplash

In 2026, 65% of people aged 14 and older in the Netherlands have used AI tools such as ChatGPT at some point, and 40% use them at least weekly (Ruigrok 2026). At the same time, about a quarter of Dutch residents aged 15 or older mainly speak something other than Dutch at home: a regional language, a dialect, or another language such as English, Polish, or Turkish (CBS, 2022).

Yet we know very little about how they bring their multilingual repertoires to these tools. Do they prompt in Dutch, in English, in a heritage language, or in a mix? When do they switch, and why? This project maps multilingual ChatGPT use in the Netherlands through a survey, followed by interviews with selected participants. It establishes patterns of who uses which languages with AI and for which purposes, and examines which ideas about ‘good’ or ‘proper’ language shape these choices. The project brings a linguistic perspective to public and academic debates on AI in multilingual societies.

Sources:

This website uses cookies.  More information.