Pleased graduates in Leipzig.
Image: dpa
Artificial intelligence is questioning how we measure academic performance. Instead of a panicked debate about abuse, it would be better to adapt to the needs of the third millennium. A guest post.
Dhe discussion about the meaning, content and significance of examinations is a constant companion in the education debate. New Artificial Intelligence (AI) like the text synthesis system ChatGPT are now fundamentally questioning the way we conduct exams: If a machine can pass exams or provide essential parts of a correct answer, are these exams still up-to-date and appropriate? Even if the public discussion is currently more about the problem of abuse, the question of the nature of exams seems to us to be more decisive overall. We think this is a very welcome question as it gives ChatGPT the role of a catalyst. ChatGPT is more of an opportunity than a problem – and time is of the essence.
We perceive a sometimes undifferentiated, sometimes panicked discussion in which responsibilities are shifted back and forth between those involved. We suggest – initially focusing on the tertiary sector – to start a process now in which universities, faculty days, professional associations and ministries in particular use the unique opportunity to adapt examinations to the requirements of the third millennium. ChatGPT is in the world and is getting more and more impressive with the new underlying model GPT-4 that the American company Open AI has just released, which can also handle inputs other than text. It is up to us to use the technology sensibly and bring about the long-awaited paradigm shift in the auditing system.