ki-in-der-bildung-640chenspec @pixabay

Application potentials and challenges of artificial intelligence in education

sprungmarken_marker_1780

Background and central aspects of the topic

In the future, artificial intelligence (AI) will likely play a much more vital role in the field of education, among other fields. Thus, in the context of the COVID 19 pandemic, teaching has shifted to digital platforms. For example, pupils and students are now working with personalised recommendation systems and AI-based learning apps. It is to be expected that AI will gain importance in the future not only as a subject of education in the school, tertiary and vocational sectors, but also as an educational medium.

At the learner level, possible applications include e. g. the support of a (digitally organised) learning process by means of evaluating available data (learning and data analytics). By collecting data on the learning progress (process data from digital learning platforms, but also data obtained by means of special sensors such as eye-tracking glasses), it is possible to provide hints with regard to strengths and weaknesses in knowledge acquisition or to offer the respective appropriate learning content. Moreover, intelligent tools for creating augmented or virtual realities (AR/VR) can provide a learning experience that is not subject to the limitations of analogue learning (e. g. in scientific experiments or simulations). For teachers, AI systems offer manifold possibilities to adapt the teaching of content and learning of competences more closely to the respective learning group and to individual learners, for example if the (automated) evaluation of the collected data reveals a need for support. At the institutional level, the data obtained can be used to prospectively assess learners' chances of success and, if necessary, to provide appropriate support.

So far, in Germany, AI in education is still in the early stages of development and implementation – both with regard to AI systems developed specifically for education and to the use of generic AI applications (e. g. language processing) in educational contexts. Furthermore, legal, pedagogical as well as ethical questions about the collection and use of learners' data and the risks of applying AI in learning processes still remain widely unanswered.

Objectives and approach

The tangible effects of AI applications on learners, teaching-learning processes or their organisation can only be analysed when a larger number of application examples is available and the results of current and future research projects have been evaluated.

This is why it has to be examined first – by way of a monitoring study – which education-related AI applications currently exist or are in development in Germany. In addition, the state of knowledge regarding the application potentials and challenges of AI applications in education as well as regarding possible opportunities for improving educational processes shall be evaluated and summarised in the form of a synopsis. Based on interviews with experts (from the fields of AI and education), a roadmap of expected developments should then be drawn up and possible impacts should be identified. For a concentrated analysis, we chose to focus on the area of ​​vocational education and training.

Project progress

Two expert reports were commissioned in October 2021 to review the status of research, development and application of AI in education. In the first on, the international academic debate is evaluated with the help of a literature analysis, among other things. In addition to an overview of the various fields of application and systems, its aim is also to identify the potential, but also possible obstacles, and to gain insights into the significance of ethical guidelines and funding programmes. The second deals with the situation in Germany. The report describes which AI-supported applications are already being used in the various areas of education, what expectations and fears are associated with them and what potential exists specifically in the area of competence-oriented learning in vocational education and training.

The expert reports are currently being supplemented by further research to take account of current developments. The final report will be submitted to the responsible TA rapporteur group of the parliamentary groups for approval in spring 2024.

In-depth investigation of ChatGPT and comparable systems

With the presentation of ChatGPT by the company OpenAI in November 2022, the possibilities of natural language processing with AI systems became visible and tangible for a broad public. The applications already discussed, especially in education and science, challenge existing practices of learning and knowledge processing, but also hold opportunities. In order to respond to this new development, the Committee on Education, Research and Technology Assessment commissioned TAB in February 2023 to investigate ChatGPT and comparable computer models for language processing in greater depth and to analyse their implications in technical, social and ethical terms. A special focus will be on educational processes in schools, higher education and training and further education. The results published as background paper, were presented and discussed at a  public expert meeting of the Committee on 26 April 2023.

Hintergrundbild: Auf einer schwarzen Tastatur am unteren Bildrand liegt ein aufgeschlagenes Buch, darüber schweben rote Wörter in Form von Etiketten, über denen der Buchstabe A schwebt Teresa Berndtsson / Better Images of AI / Letter Word Text Taxonomy / Licenced by CC-BY 4.0.
Opportunities and risks of AI applications

 

Expert discussion in the Education Committee on Wednesday, 26 April 2023

Nachschauen

Publication


In the Bundestag

(only in German)

Publications in context (only in German)


Presentations
ChatGPT - Auswirkungen auf Gesellschaft und Politik
Albrecht, S.
2023. FachMittag zum Thema ChatGPT und Auswirkungen auf Gesellschaft und Politik (2023), Vienna, Austria, June 2, 2023 
ChatGPT und der wissenschaftliche Text. Eine Annäherung aus der Perspektive der Technikfolgenabschätzung (und eigener Betroffenheit)
Albrecht, S.
2023. Interdisziplinäre Tagung KI – Text und Geltung: Wie verändern KI-Textgeneratoren wissenschaftliche Diskurse? (2023), Darmstadt, Germany, August 25–26, 2023 
Sprengt ChatGPT das Bildungssystem? Generative KI als Potenzial und Herausforderung
Albrecht, S.
2023. Fachkonferenz "Generative KI": „ChatGPT und Generative Künstliche Intelligenz: Quantensprung oder leeres Versprechen?“ (2023), Munich, Germany, November 30, 2023 
Cover TKP Nr. 52, Sprich mit mir! Perspektiven für den Einsatz KI-basierter Dialogsysteme

Themenkurzprofil Nr. 52

 

Sprich mit mir! Perspektiven für den Einsatz KI-basierter Dialogsysteme 
Peters, R.
2022. Büro für Technikfolgen-Abschätzung beim Deutschen Bundestag (TAB).
doi:10.5445/IR/1000143462

Cover: TKP Nr. 42: Learning Analytics – Potenzial von KI-Systemen für Lehrende und Lernende

Themenkurzprofil Nr. 41

 

Learning Analytics – Potenzial von KI-Systemen für Lehrende und Lernende.
Peters, R.; Bovenschulte, M.
2021. Büro für Technikfolgen-Abschätzung beim Deutschen Bundestag (TAB).
doi:10.5445/IR/1000131769