|ChatGPT and other computational models of language processing - basics, applications and possible implications|
Rarely has a computer system attracted as much attention and debate around the world as ChatGPT has since its launch in November 2022. The chatbot is based on a computer model that has been trained to process linguistic data using artificial intelligence (AI) methods. It can generate eloquent responses on a wide range of topics in a very short time, create entire essays or computer programmes, and use language styles such as poetry, jokes or discussions in different languages. On the one hand, the high public profile raises expectations about the potential applications. On the other hand, it can also obscure a realistic assessment of the possibilities and limitations of such systems, as well as their social impact. In order to provide guidance for the ongoing debate, this background paper examines - the technological developments on which the system is based, - the the potential and limitations of the technology - possible applications, in particular in the field of education and - the possible impacts of an application. The aim of the background paper commissioned by the Committee on Education, Research and Technology Assessment in February 2023, which is the basis for a public expert discussion in the committee, is to compile sound information on these aspects and to identify questions under which the role of language-processing computer models can be further monitored and discussed.