Skip to main content
Course overview
Skip to - Close

1.4.2 Open-Source-Interface HAWKI

1.4.2 Open-Source-Interface HAWKI

Completion requirements
View

Privacy-friendly, transparent, independent: HAWKI demonstrates how AI can be specifically designed for higher education institutions.

HAWKI, developed at Hochschule für angewandte Wissenschaft und Kunst Hildesheim/Holzminden/Göttingen (HAWK), serves as an excellent example of how open-source AI technology can be used in the higher ed sector in a practical and privacy-compliant manner. For example, UHHGPT at Universität Hamburg and the chatbot “TUKI” at TU Hamburg are based on HAWKI and are available to all members of the university. Hochschule für Musik und Theater Hamburg has its own chatbot “Ask Wolfgang”.

 

What is HAWKI?

HAWKI is a campus-owned, data protection-compliant interface for generative AI, developed specifically for use in a higher education context. What makes it special is that students and staff do not need a separate account; they can simply log in using their regular university credentials. An important aspect here is data protection – no personal data is stored or passed on to third parties.

HAWKI was developed in the HAWK Interaction Design Lab and has several objectives:

  • to give all campus members the opportunity to integrate AI into their work processes
  • to create a meeting place where new ways of working can emerge
  • to promote an internal discussion on the use of AI

 

HAWKI2: The next stage of development

A new version, HAWKI2, has been online since February 2025. The new version places particular emphasis on the following:

Collaboration via chat rooms

Campus members can exchange ideas in interactive chat rooms – similar to messaging services – and involve the generative AI in the conversation at any time by addressing it as ‘@hawki’. This makes AI a participant in group discussions, enabling joint learning and collaborative work with AI support.

Transparency for academic work

An innovative feature is the automated documentation of prompts and an intelligent export function with a summary. This creates transparency in AI-supported work processes – particularly important for use in examinations and academic work, where the traceability of results is crucial.

Flexibility through modular architecture

HAWKI2 was designed with a modular architecture that enables the rapid and flexible integration of new features. This allows educational institutions using HAWKI to tailor the interface specifically to their specific needs. The HAWK plans to add new features at regular intervals.

Diversity through various model options

A particularly interesting feature is the ability to use different language models. This opens HAWKI up to intercultural perspectives and a wider range of use cases in higher education.

The differently trained models are also ideally suited for critical engagement with the values, norms, cultural influences and potential biases contained within the respective models.

 

A dedicated AI system for higher ed institutions: data control and ethical guidelines

HAWKI deliberately focuses on cross-institutional collaboration, thereby differing fundamentally from commercial AI applications. At its core lies the digital sovereignty of the higher ed institutions. Instead of passing on data from students and researchers to commercial providers, the institution retains control.

The aim is to create a networked ecosystem that enables campus members to interact with generative AI in their own way – without rigid constraints, but with clear ethical and pedagogical guidelines.

This is in line with the fundamental principle of open source: technology that can be adapted to one’s own needs and is not restricted by commercial interests.

  

HAWKI as a bridge between theory and practice

The use of HAWKI exemplifies how open-source LLMs can be utilised in practice to meet specific requirements. Whilst closed systems such as ChatGPT often dominate the commercial sector, HAWKI demonstrates that open systems offer decisive advantages, particularly in the education sector:

  1. Data protection: Sensitive learning and research data remains within the institution
  2. Adaptability: The interface can be specifically optimised for teaching purposes
  3. Transparency: Students can understand how the AI arrives at its answers
  4. Learning effect: Working with different models sharpens understanding of AI systems

This combination makes HAWKI a valuable example of the practical application of open-source AI technologies in an educational context and demonstrates how theoretical concepts can be put into practice.

 

 

  

 Information

 

💡 Learning Summary Chapter 1.4.2: Open-Source-Interface HAWKI

  • HAWKI is a data protection-compliant AI interface for universities: Developed at HAWK, HAWKI enables secure access to generative AI via university login – without external data processing or commercial providers.
  • HAWKI2 introduces new features for collaboration and transparency: With chat rooms, automated prompt documentation and a selection of different models, HAWKI2 supports collaborative and scientifically traceable work.
  • Open source creates autonomy and educational potential: Through openness, adaptability and model diversity, HAWKI is a practical example of the critical, ethical and pedagogically reflective use of AI in a university context.