Artificial Intelligence in Healthcare and Education
Abstract
The potential and challenges of using Generative artificial intelligence (GAI) in medical education are widely discussed, yet its use by medical students and faculty remains under-researched in the UK context. This exploratory sequential mixed-methods pilot study empirically investigates and compares student and faculty experiences and perceptions of using GAI in medical education. A questionnaire was developed based on focus groups and administered to undergraduate medical students and faculty within a UK medical school. Descriptive analysis and comparative analysis were used to quantify and compare student (n = 29) and faculty (n = 32) experiences and perceptions of GAI. Reflexive thematic analysis of open-ended question responses was undertaken to complement quantitative results. We found that GAI is being used for a variety of purposes by students and faculty in learning/teaching/assessment. However, both of them generally have relatively low self-confidence in GAI-related knowledge and skills. Faculty members tend to have stronger concerns of GAI limitations and its ethical challenges. Males are more confident in using GAI and have more positive attitudes towards GAI than females. The findings also provide preliminary evidence for the importance of open communication of GAI between students and faculty and the need for schools’, universities’, and national regulators’ strategic support to ensure faculty and student ethical and effective use of GAI and equitable access to GAI technology between subgroups. Our findings and the pilot tools can inform future research investigating the use of GAI in medical education in other institutional context or at a larger scale.
GAOKAO-Bench is introduced, an intuitive benchmark that employs questions from the Chinese GAOKAO examination as test samples, including both subjective and objective questions that contribute a robust evaluation benchmark for future large language models and offers valuable insights into the advantages and limitations of such models.
Xiaotian Zhang, Chun-yan Li, Yi Zong et al.· arXiv.org· 216 citations· ⚡17
Empirically, PRISM reduces the end-to-end time for data selection and model tuning to just 30% of conventional pipelines, and achieves this efficiency while simultaneously enhancing performance, surpassing models fine-tuned on the full dataset across eight multimodal and three language understanding benchmarks.
Jinhe Bi, Yifan Wang, Danqi Yan et al.· arXiv.org· 73 citations· ⚡4
The method, ECCOLA, is presented, which aims at making the high-level AI ethics principles more practical, making it possible for developers to more easily implement them in practice.
Ville Vakkuri, Kai-Kristian Kemell, P. Abrahamsson· EUROMICRO Conference on Soft...· 64 citations· ⚡6
This paper designs Markov decision processes (MDPs) for different combinatorial problems and proposes to train conditional GFlowNets to sample from the solution space and demonstrates that GFlowNet policies can efficiently find high-quality solutions.
Dinghuai Zhang, H. Dai, Esmeralda S. Whitammer et al.· Advances in Neural Informati...· 59 citations· ⚡8
An empirical study on the current state of practice in artificial intelligence ethics is conducted by means of a multiple case study of five case companies, which indicates a gap between research and practice in the area.
Ville Vakkuri, Kai-Kristian Kemell, Joni Kultanen et al.· arXiv.org· 56 citations· ⚡6