Digital autonomy is a key consideration. The UvA aims to reduce dependence on large technology providers and maintain greater control over infrastructure, conditions and data handling. With many commercial AI services, it remains unclear how user input is processed or reused. UvA AI Chat is designed so that prompts are not stored, not used to train models and remain fully anonymous. User data is not shared for commercial purposes.

Equal access is another important principle. While many commercial AI tools operate behind paywalls or offer premium features, UvA AI Chat is available free of charge to all students and staff. This ensures that access to generative AI does not depend on individual financial resources.

The chat environment is also positioned as a learning space. Students can gain experience with AI in a controlled and safe context, including understanding the boundaries between responsible use and misuse. According to Benneker, misuse begins when generated content causes harm to others. Clear safeguards allow the tool to function as a training ground for skills students will need later on.

Future development focuses on integration with existing UvA systems. Lecturers can already create course-specific personas, and planned connections with educational and administrative platforms will allow users to retrieve UvA-specific information directly through the chat.

Sustainability is explicitly part of the design choices. While generative AI is not inherently energy-efficient, the UvA aims to make energy use transparent and to offer alternatives. Users can choose between smaller, more energy-efficient models and larger models for specialised tasks. The university also collaborates with data centres that meet sustainability standards.