Humboldt-Universität zu Berlin - Artificial Intelligence at HU Berlin

Using LLMs

What are possible use cases for the application of data protection-compliant Large Language Models (LLM) in administration, research and teaching?

We would like to present some use cases for the application of Large Language Models (LLM) in various areas at HU Berlin. They are by no means all-encompassing and we are happy to receive further use cases.

Please try out the LLMs and experiment with the purposes for which you could use the models. As nothing is logged, you do not need to worry about data being passed on to third parties.



  • Support in the generation of social media posts and contributions
  • Creation of diagrams and illustrations from verbal descriptions
  • Support in revising answers to frequently asked questions
  • Support with the creation of job advertisements
  • Writing texts for invitations to events and workshops
  • Discussion of alternatives for structuring processes
  • Virtual "sounding board" for the discussion of ideas
  • Creation of summaries of longer texts
  • Inspiration for writing greetings and addresses of all kinds
  • Support in writing justifications (e.g. for the necessity of certain events, expenses, etc.)



  • Create summaries of scientific articles
  • Support with spelling and grammar of scientific articles
  • Support with alternative formulations, e.g. sentence beginnings, sentence structure, etc. for scientific articles
  • Support with the generation of code for data analyses and simulations
  • Generation of diagrams and illustrations from linguistic descriptions
  • Suggestions for the structure of presentations
  • Inspiration for general introductions to generic topics and interview requests with general or thematically generic questions
  • Planning of workshops
  • Translation of research reports into different languages
  • Recommendation of suitable tools/libraries for specific activities
  • Support with various forms of scientific text analysis, including in the field of NLP (Natural Language Processing), but also for questions of historical contextualization
  • Support with the translation of historical language forms (e.g. Latin, Ancient Greek, Middle High German) (for models re-trained on the corpus in question) natural language and complex queries on a very large text corpus



  • Generation of exercise questions and test tasks
  • Support with the creation of lecture notes and summaries
  • Support with the creation of lecture slides
  • Help with writing lecture announcements and course descriptions
  • Support with translations of teaching materials
  • Creation of diagrams and illustrations from verbal descriptions
  • Support in structuring lecture topics
  • Writing reflection tasks
  • Preparing teaching materials for inclusive education and special needs
  • Preparing of materials in easy language
  • Help with the preparation of summaries for seminar papers
  • "Sparring partner" for learning, e.g. programming languages
  • Live demo for code generation with subsequent critical validation of the code
  • Dialogue partner for learning foreign languages (different language registers, different language levels)