Generative AI in teaching

You have not yet dealt with Generative AI

Then we recommend starting with: "ChatGPT for non-computer scientists: Keys to understanding artificial intelligence and its applications in university teaching (dghd series "KI in der Hochschullehre)".

Want to try out Generative AI for yourself?

Currently, H-BRS is preparing the data protection-compliant provision of OpenAI's large language models via an in-house hosted interface to anonymise your data. For this purpose, H-BRS will book a credit that employees and students can use with their H-BRS access data. It will also be possible to upload files. The provision should take place in Q4/2024. Current information on the provision will be published on this page as soon as it is available.

Until the internal solution is available, you have the opportunity to try out external models:

• The GWDG offers access to various models via the AcademicCloud of the State of Lower Saxony (https://chat-ai.academiccloud.de/). The access is possible via the H-BRS access data (MIA) and TXT files can be uploaded as context. Information on the offer can be found on the pages of the University of Göttingen (https://www.uni-goettingen.de/de/686446.html).
• You can register for a free version at ChatGPT (https://chat.openai.com/). You can borrow a notebook with ChatGPTPlus at the reception desk of the Sankt Augustin library to try out the paid version without entering any personal data.
• A comparable offer to ChatGPT is provided by the French software company Mistral AI with the service Le Chat, which can be found here (https://chat.mistral.ai).
• Bing's CoPilot is also an interface to Generative AI and can be accessed in the Microsoft Edge browser via the blue loop at the top right. The web, a website or a subpage can be selected as the data source. Here (after application) it is also possible to generate images, for example, in a simple way. Commands would be, for example: "Create the image of a flower meadow", "Add a wind turbine".

Didactics on teaching with and despite AI

For an overview, we also recommend the information provided by the Hochschulforum Digitalisierung: https://hochschulforumdigitalisierung.de/chatgpt-in-studium-und-lehre/

The ZIEL will be offering further workshops on the topic of AI in the winter semester. You will gradually find details on our workshop schedule (https://www.h-brs.de/de/ziel/workshop-termine). You will receive the corresponding invitations by email approx. 6 weeks before the respective workshop date.

Suggestions for AI newsletters

• Do you also struggle to keep up with all the important information on the rapid development of AI? Maybe there's something for you in the "8 AI newsletters you should subscribe to". https://blog.hubspot.de/marketing/ki-newsletter
• You can also find a newsletter focussing on AI and education here: https://www.unidigital.news/. You can subscribe to the newsletter via LinkedIn: https://www.linkedin.com/newsletters/ki-bildung-7088522528528195584/

Want to know more and delve deeper into the topic?

Then we recommend the AI Campus (https://ki-campus.org/), e.g.

Self-study module (beginners without prior knowledge)

KI-Campus "Introduction to AI"

8 weeks of 5 hours each, pass/fail assessment, free

On completion of the course you will be able to..

• Realistically categorise headlines and news about AI from the internet and press.
• differentiate between important terms in machine learning and understand how it works.
• Critically evaluate the benefits and risks of AI and categorise the current debate on the regulation of AI.
• to assess the future of society with AI and to understand the impact of AI on the personal, professional and public environment.

Self-study modules with a technical component (beginners with no prior knowledge)

KI-Campus: "AI for everyone: Introduction to Artificial Intelligence"

14 weeks of 3 hours each, certificate of attendance, free

On completion of the course you will be able to...

• Explain basic methods of Artificial Intelligence and simple application examples, as well as what is and what is not currently state of the art in Artificial Intelligence.
• Describe different types of data as well as possible pitfalls and problems of data in the context of artificial intelligence.
• Execute simple operations and basic commands in Python.
• Explain various ethical and legal aspects and challenges of artificial intelligence.

Also recommended is the expert report"Didactic and legal perspectives on AI-supported writing in higher education". This report consists of a didactic and a legal section. In the didactic section, under 3., reference is first made to possible new learning objectives in academic writing. Furthermore, concrete solutions are offered for dealing with exams without supervision. The legal section is also well worth reading with regard to copyright, scientific misconduct, etc.

Exams and AI

There are helpful websites on the topic of generative AI and exams, particularly in English-speaking countries. We recommend the website of Melbourne University, which deals with the topic of Assessment and AI from a practical perspective and provides recommendations for action with concrete examples from examination practice. Two areas are covered here: Designing assessment tasks that are less vulnerable to AI and Using AI to enhance assessment. The following review provides an overview of publicly accessible guidelines from 50 different universities on how to deal with AI in assessment design. The following topics are covered in the guidelines: academic integrity, advice on assessment design and feedback culture. It analyses the extent to which the various higher education institutions have adapted their assessment guidelines to take account of the use of Generative AI. The QAA (Quality Assurance Agency for Higher Education/UK) provides recommendations for the development of new examination strategies and compares different types of examination in terms of their strengths and weaknesses. A suitable AI Assessment Flowchart can be found on the following website. In addition, the handouts from the University of Stuttgart provide a deeper insight into AI tools and exams, especially for unproctored, written exams such as bachelor's and master's theses.

If you have any questions, please contact us!