Improving Remote Examination Formats : Addressing Challenges and Providing Recommendations for University Educators in the Age of ChatGPT

• The availability of Large Language Models (LLM), such as the widely accessible ChatGPT, has raised concerns regarding the reliability of assessment formats in university education. This presentation sheds light on the potential impact of ChatGPT on existing remote examination formats and offers recommendations to overcome associated challenges. Through interviews with three computer science educators and an analysis of insights from blog posts, social media platforms, and academic forums, we gathered diverse opinions and experiences related to ChatGPT's impact on university assessment formats. Our findings reveal challenges, particularly in cheating prevention, prompting the need for educators to adapt their teaching methods to this evolving landscape. While GPT performs well in answering open-ended questions, it struggles with calculations and problem-solving, with multiple-choice questions posing additional challenges. Incorporating interactive activities and oral examinations should be prioritized in this context. As ChatGPT gains traction, institutions must address concerns to maintain academic integrity.

Ämnesord

SAMHÄLLSVETENSKAP  -- Utbildningsvetenskap (hsv//swe)

SOCIAL SCIENCES  -- Educational Sciences (hsv//eng)

Nyckelord

Large language models

LLM

ChatGPT

GPT

assessment formats

Publikations- och innehållstyp

ref (ämneskategori)

kon (ämneskategori)