Με τη χρηματοδότηση της Ευρωπαϊκής Ένωσης. Οι απόψεις και οι γνώμες που διατυπώνονται εκφράζουν αποκλειστικά τις απόψεις των συντακτών και δεν αντιπροσωπεύουν κατ'ανάγκη τις απόψεις της Ευρωπαϊκής Ένωσης ή του Ευρωπαϊκού Εκτελεστικού Οργανισμού Εκπαίδευσης και Πολιτισμού (EACEA). Η Ευρωπαϊκή Ένωση και ο EACEA δεν μπορούν να θεωρηθούν υπεύθυνοι για τις εκφραζόμενες απόψεις. Lessons learned N/A Implications for practice The study highlights the potential benefits and limitations of current AI approaches in the classroom, including the ethical implications, linguistic, and cultural contexts. It is recommended that future research should prioritise promoting AI literacy, developing ethical AI guidelines and policies, and identifying and addressing biases in AI algorithms and models for equitable and responsible AI practices in higher education institutions. Μελέτη περίπτωσης 22: Θα περάσει το ChatGPT τα διαδικτυακά κουίζ; Προσαρμογή μιας στρατηγικής αξιολόγησης στην εποχή της γενετικής ΤΝ. General information Reference/Source: Raftery, D. (2023). Will ChatGPT pass the online quizzes? Adapting an assessment strategy in the age of generative AI. Irish Journal of Technology Enhanced Learning, 7(1). https://doi.org/10.22554/ijtel.v7i1.114 University/HE institution: South East Technological University Course and subject domain: Quantitative techniques modules for first-year business students. Aim: To investigate the accuracy of using ChatGPT to answer online quiz questions and explore students' attitudes and experiences with online quizzes. Target group: First-year business students who are taking quantitative techniques modules Description of case AI tool used: ChatGPT versions 3.5 and 4 : https://chat.openai.com/ A detailed description of what happened: Researchers copied and pasted each question from the quizzes into ChatGPT, and then pasted the response back into the virtual learning environment (VLE). This was done initially for each of the quizzes using ChatGPT-3.5 and then later with updated versions of ChatGPT. Lessons learned ChatGPT has the potential to answer multiple-choice questions, attaining correct answers on average for 50% of the attempted questions using ChatGPT-4 compared to 43% with ChatGPT-3.5. Implications for practice Future research is recommended to focus on improving ChatGPT's reliability and accuracy while maintaining the academic integrity of students.
RkJQdWJsaXNoZXIy NzYwNDE=