Freedom of choice

The ability to switch between different models contributes to the digital autonomy of the UvA and, by extension, to yours as a user. As a user, you can compare different models and decide for yourself which model best matches your needs or tasks. 

Differences in use

At first glance, using Claude is the same as using gpt-5.1 or the other models. The output, however, may differ. Claude is known for handling long documents, content-related uncertainties (with a lower chance of hallucinations), and bias better than many other models. 

By contrast, gpt-5.1 has a more pragmatic and less cautious approach. 

Default model

To use the Claude models instead of the system default, gpt-5.1, go to settings > general > default model in UvA AI Chat. You can switch between the models as often as you like, depending on your needs. 

Your privacy and data remain protected

Whichever model you use in UvA AI Chat, the following always applies: data processed via UvA AI Chat, including chat content, is never accessible to parties outside the UvA and is not used to train models. Regardless of the model chosen, all data is anonymised before being sent, so it cannot be linked to a specific person. 

If you delete a chat, it is truly deleted. 

Disclaimer: risks, plagiarism and fraud

Curious about how, when and why you are allowed to use UvA AI Chat to support your studies? Before using it, please visit this page: 
https://student.uva.nl/en/information/uva-ai-chat to learn more about the possibilities. 

Questions?

Email AIinhetonderwijs@uva.nlExternal link to get an answer from our AI Living LabExternal link.