AG: Generative KI

Generative AI in Studies and Teaching

With the introduction of ChatGPT in November 2022, the topic of Artificial Intelligence (AI) has gained widespread attention, and the transformative potential of the technology is unmistakable – including in academic work and educational processes. But what exactly does generative Artificial Intelligence (GAI) mean? What applications exist beyond ChatGPT, and what ethical, legal, and didactic questions does its use in university teaching raise? The “Working Group on Generative AI in Higher Education”, initiated by the university’s presidium, is dedicated to addressing these and other questions.

About the WG:GAI

The aim of the working group is to provide a central point of contact for this very diverse and comprehensive topic, in order to find answers to these partly unresolved questions together with as many university stakeholders as possible and to pass them on in the form of concrete offerings. These include, for example, specific recommendations for action for teachers and students or exclusive access to various AI tools and their application scenarios. The exchange about the use of these offerings and their joint further development takes place in a university-wide Community of Practice (CoP), which meets regularly at various events and workshops or exchanges asynchronously via newsletters and platforms such as our Discourse server.

The working group, led by Dr. David Weiß (studiumdigitale, Virtual Teaching WG) and Julia Schmitt (studiumdigitale, Media Didactics), is composed of representatives from the departments, specialist centers, and experts for study and examination regulations, as well as student members.

Previous Activities

Handouts and Recommendations for Action

As one of its first activities, the WG: GAI, together with teachers and students, has created handouts on dealing with generative AI in studies and teaching. The aim was to create comprehensive guidance for both students and teachers as quickly as possible with the increasing emergence of AI tools, even though legal, didactic, and technical questions are still open. For this reason, the handouts are an article in the ‘Lehre virtuell’ knowledge database, which is constantly updated.

Handouts on the ‘Lehre virtuell’ Page

General Introduction for Everyone

  • Recommendations and information on intellectual property, personal data, currentness, and critical reflection
  • What is Generative Artificial Intelligence (GAI)?
  • What applications and services are available?
  • Use of Generative AI in Studies and Teaching
  • Frequently Asked Questions (FAQ)

Target Group Specific

In addition to the general introduction to the topic of generative AI, you can find two further articles on the ‘Lehre virtuell’ page specifically for the target group teachers and the target group students.

  • How can AI change teaching and could it be used?
  • What should be considered when using AI in studies?
  • How can AI support learning and study organization?
  • How can AI support writing processes?
  • Good scientific practice such as labeling requirements and declaration of originality
  • Information on dealing with suspected cases of unauthorized use of AI tools in written works
Recommendations for Action for Teachers
Recommendations for Action for Students

Community of Practice

In addition to the general handouts and recommendations, another task of the WG: GAI is to identify relevant use cases together with teachers and students and to test them with a coordinated development of an AI infrastructure (tools + qualification) at Goethe University and to disseminate them widely. This is the Community of Practice for Generative AI (CoP-GAI).

Source: Wenger-Trayner et al., 2022

Domain:

Application, reflection, and integration of Generative AI in the context of teaching and learning.

Community

More experienced teachers (early adopters) from Goethe University from different disciplines and with as diverse teaching/learning concepts as possible; selected students from higher semesters.

Practice

Access to tools, their specific use in teaching, and the exchange of experiences. Technical and didactic maturation of scenarios for translation into central services of the university (SMS) and use for less experienced users (onboarding, awareness-raising, best practices).



Join Community of Practice

AI-ToolLab for Use in Teaching

One of the biggest challenges in using generative AI in studies and teaching is the selection and use of available tools. Currently, the greatest potential lies in the rapidly developing free market (large number of innovative tools – mostly free to start). But it also comes with the greatest risks (copyright infringement, lack of data protection, no equal opportunities – better models only available for a fee). For this reason, studiumdigitale, the innovation unit for digital teaching and learning at Goethe University, provides an AI-ToolLab. The task of the Community of Practice, initiated by the AG:Generative AI, is to initially test these tools and, together with the AG, bring them to maturity until they can be made available as a service to the entire university.

Goals and Challenges

  • Easy access to innovative tools
  • Opportunity to gain initial experience in using GPT-x/image generators in one’s own context
  • Aligned with qualification offers, how-tos, and tutorials
  • Scouting for data protection and copyright-compliant solutions in the context of generative AIs
Open AI-ToolLab

FAQs

How can one ensure that all students/teachers have access to a tool?

As long as there are no corresponding offers from the university itself, the decision lies with each individual, which means access cannot be guaranteed. The university is currently working on appropriate solutions, but this is delayed due to various aspects (lack of data protection, insufficient quality of results, rapid development, high resource consumption, diverse requirements). As soon as a solution exists, this will be communicated via the newsletter and the usual channels. If you have any needs or suggestions regarding specific tools, please let us know via ag-gki@studiumdigitale.uni-frankfurt.de, and we will incorporate this into the scouting process.

What to do if students/teachers do not want to create an account with ChatGPT, for example?

Alternative tasks/materials should always be offered that can be completed without AI tools or ChatGPT. Learners/teachers should not be disadvantaged if they do not want to or cannot create an account. In teaching contexts, group work can help ensure that even people without their own access can try out the use or complete tasks.

Are there guidelines and conditions for correcting and grading papers/written exams?

Generative AI can assist in evaluating practice assignments or student papers. However, it is important to apply one’s own assessment criteria and critically question the AI’s results to avoid systematic unequal treatment of students (see Where can I use AI in teaching?). When it comes to evaluating work where AI is used, it is advisable to clearly communicate in advance what is allowed and under what conditions (see Good Scientific Practice, Labeling Requirements and Declaration of Independence and What should I tell my students about using AI?).

Can AI tools or categories of AI tools be banned for one’s own teaching? If yes, how?

Fundamentally, it is the responsibility of the teachers to decide whether and for what purpose AI tools may be used. However, issuing bans can always lead to non-compliance. A transparent approach to tools and clear rules in the classroom context ensure a trusting, creative, and constructive interaction with one another. It is always important to question what learning objective should be achieved and to what extent the use of AI tools could impair this. . As an alternative to banning tools, adapting one’s own teaching or the assigned tasks could be a solution to address this.

Can I (have) AI-generated texts detected software-assisted?

Proving deception when using generative AI tools is currently still extremely difficult, as no reliable detection software is available to date, making AI use not easily verifiable. While there are already programs on the market such as AI Text Classifier or ChatGPT Zero. However, these are very error-prone and alone would not yet constitute sufficient evidence of deception.

Is there a way to try out AI tools without registering?

It depends on the service and the AI tool. Some services offer the possibility to try out a tool without restrictions, possibly even without registration. Often, however, registration is mandatory, or usage is only possible through payment. Experience shows that the requirements for use decrease with the development stage and “professionalism” of the services. Experimental offerings are usually free and sometimes testable without registration. Large, finished products with extensive functionality usually cost something and require at least registration. This is because providing AI services represents a high financial expense for the operators. AI servers with the required computing power are expensive, and these costs must be recouped by the providers or usage must be limited for free offerings. Unfortunately, this is effectively done by requiring users to register.

Are there courses on dealing with AI?

For teachers at Goethe University, there are various offerings from IKH, the Writing Center, and studiumdigitale. You can find the offerings under Use of Generative AI in Studies and Teaching. You can find additional very good, free self-learning offerings outside the university, for example, at https://ki-campus.org/.

How do I cite correctly or how do I properly indicate the use of AIs?

You can find a possible suggestion for citing AI-generated texts under Good Scientific Practice and Labeling Requirements and Declaration of Originality.

Contact and Newsletter

Questions? Feel Free to Contact Us!

Or stay up to date by subscribing to our newsletter.

Contact via email
Subscribe to the newsletter