Dear members of the University, The text-generating artificia | دکتر روشنی متخصص کاربرد پرتوها

Dear members of the University,

The text-generating artificial intelligence (AI) Chat-GPT is capable of writing even longer specialized texts on scientific topics, which are often indistinguishable from written texts by human experts. It was also shown that Chat-GPT could even successfully complete parts of final exams at universities.

This technological development holds both opportunities and risks for research and teaching at universities. On the one hand, linguistically mature repetitive research work, summaries or utility texts can be efficiently created by generative AI in the future. On the other hand, classic asynchronous examination formats such as term papers may also become obsolete because students can use text-generating AIs to write their papers without having to make the required personal contribution.

As things stand, however, the undeclared use of AI-generated texts in term papers and theses contradicts the declaration of independence to be submitted by students and would thus be tantamount to cheating, which would be sanctioned if discovered.

However, concrete proof of AI use is difficult because common tools for detecting plagiarism by comparing texts will no longer suffice. Although the first tools for detecting AI-generated texts exist, these at best indicate the probability that a text was AI-generated and, at least at present, are unlikely to stand up to legal challenge.

This technological development challenges universities to come to an understanding about the limits and scope of the integration of AI into learning and teaching. This process includes, in particular, a reflection on future-oriented examination formats and the competences to be taught in dealing with AI.

Moreover, powerful and flexible AIs also raise fundamental social questions about the displacement of human paid labour and the necessary adaptation of affected life models. The transparency, verifiability, and non-discriminatory nature of the influence of AI decisions and communication on human life require new ethical and legal norms.

It is also clear that the impressive performance of Chat-GPT is only the beginning of an accelerated development of ever more powerful AIs. We expect AIs that will match or even exceed human performance in more and more areas of life. A simple reflective discussion of the status quo is therefore not sufficient.

With this in mind, the University of Jena will initiate a participatory process of debate on the use of text-generating artificial intelligences in the coming months. Further education and regular training of students, lecturers and employees on the basics, skills and responsible use of AIs must form the basis for future discourse at the University of Jena. With the participation of students and employees, we will be developing recommendations for action in a timely manner.

Best regards,

The Executive Board of Friedrich Schiller University Jena