open search

Notes on the use of genAI4BUW at the university of Wuppertal

Legal information

In particular, the University of Wuppertal enables the use of the chatbot ChatGPT from OpenAI via a central interface. This allows employees and students of the University of Wuppertal to use the chatbot via their ZIM login and no personal traffic data is transmitted to the OpenAI servers. genAI4BUW is a service offered by ZIM.

  • These instructions for use and the terms of use of OpenAI and Microsoft must be read carefully before using genAI4BUW and observed during use.
  • No copyrighted material may be used as input for genAI4BUW if the rights to this material are not owned by the user. This applies to the input of images, copies of texts, examination results, open access publications, etc. if applicable. If such material is to be used, its lawful use must be ensured (e.g. by obtaining the author's consent) and must be documented and stored in writing.
  • No personal data of third parties pursuant to Art. 4 para. 1 GDPR may be used as input for genAI4BUW. This includes real names, contact details, e-mail addresses, nationalities, age, marital status, photo, audio or video recordings.
  • No particularly sensitive personal data of third parties pursuant to Art. 9 para. 1 GDPR may be used as input for genAI4BUW. This includes ethnic origin, political opinions, religious or philosophical beliefs or trade union membership as well as the processing of genetic or biometric data for the purpose of uniquely identifying a natural person, health data or data concerning a natural person's sex life or sexual orientation.
  • If personal data of third parties is used when working with genAI4BUW, it must be anonymised before use so that no conclusions can be drawn about the person. It is also recommended that you do not enter your own personal data, as the input data is processed on servers outside the University of Wuppertal.
  • genAI4BUW and other AI systems may not be used (e.g. in the context of examinations) for automated decision-making in accordance with Art. 22 para. 2 GDPR. In borderline cases, comprehensible documentation must be prepared. The documentation must show that the final decision is made by a natural person.
  • When writing the input, as well as when using genAI4BUW in general, applicable employment contract agreements, study-related agreements, legal regulations and internal BUW guidelines must be observed to the best of one's knowledge and belief. In particular, these are agreements on confidentiality, the protection of official secrets, legal requirements or requirements from statutes and regulations of the University of Wuppertal.
  • No materials from confidential sources may be used as input.
  • If the output of genAI4BUW is to be published or distributed, it must be checked for copyright-protected content, which can arise, for example, if the AI produces similarities to existing works.
  • If the output of genAI4BUW is to be published or disseminated, this must be done in accordance with the principles of good scientific practice.
  • Please also note the information on disseminating the output of OpenAI.
  • Students must observe the applicable examination regulations and the requirements of their lecturers. In examinations, students may not use genAI4BUW at all or only within the framework of and in compliance with the examination regulations and requirements applicable to them.

This information merely provides a general overview of the legal requirements for the use of generative AI. It does not constitute legal advice and is not a substitute for seeking legal advice for specific individual cases. Further information can be found here:

Notes on good scientific practice

  • Please note that transparency is an important value in science and that the presentation of a research process is part of good scientific practice. This also applies when generative models are used to gain knowledge.
  • Texts produced by a generative model are generally not classic plagiarism due to the way the models work. Nevertheless, the use of generated texts can constitute scientific misconduct if you claim authorship of a text in which you were not significantly involved. It can also happen that texts are generated by a model that are plagiarised due to their wording because they correspond to texts that have already been published. It is your responsibility to check this.
  • Please note that generative models can produce incorrect information as well as stereotyping or discriminatory content due to biases. The responsibility for a text always lies with you if you publish it or pass it on in any other way.
  • For further information on generative models and good scientific practice, please also refer to this DFG statement.

Further information on use

ChatGPT can be helpful for various purposes. However, below you will find some tips that you should take into account when deciding whether or not to use genAI4BUW for your specific purpose:

  • From an environmental perspective, there are currently many concerns about how energy and resource intensive both training and using large language models such as ChatGPT is.
  • You should also be aware that OpenAI has used content from the internet for training ChatGPT without the consent of the respective creators, as well as copyrighted material for which OpenAI did not have rights.
  • For a product like ChatGPT to work well, a great deal of human activity is required - even if the term 'artificial intelligence' sometimes suggests otherwise. In addition to the human-generated content that has been incorporated into the training of the model, this includes the work of so-called click workers who sort, classify and evaluate training material. This often happens under precarious conditions.