Speaker

• Shlomo Berkovsky, Mcquarie University, Sidney (Australia)

Scientific Coordinator: Massimo Zancanaro

Abstract

Doctors are among the primary users and curators of electronic health records, highlighting the need for technologies supporting record keeping. Recent advancements in AI facilitate the development of systems, automating some time-consuming record keeping tasks. However, it remains unclear what AI tasks would benefit doctors most, what features such systems should exhibit, and how doctors will interact with them. We conducted semi-structured interviews uncovering the views and attitudes of doctors toward text automation with AI. The main emerging theme was doctor-AI collaboration, establishing a constructive synergistic relationship between the doctor and AI. We also developed a high-fidelity prototype of three common medical documentation tasks, Information Extraction, Summarization, and Speech-to-Text, at three distinct levels of user supervision. A new cohort of doctors interacted with the prototype and reported on relevance, perceived importance, desired automation level, and satisfaction. Doctors generally welcomed integration of AI, particularly to streamline documentation and enhance record keeping. Medium automation level was preferred, viewed as "safe with caution". Addressing safety concerns, thoroughly trialing the technology, and mitigating possible biases and medico-legal challenges are vital steps to ensure effective integration of AI into general practice.

References

Fraile Navarro, D., Kocaballi, A. B., Dras, M., & Berkovsky, S. (2023). Collaboration, not confrontation: Understanding general practitioners’ attitudes towards natural language and text automation in clinical practice. ACM Transactions on Computer-Human Interaction, 30(2), 1-34.

Fraile Navarro, D., Coiera, E., Hambly, T. W., Triplett, Z., Asif, N., Susanto, A., ... & Berkovsky, S. (2025). Expert evaluation of large language models for clinical dialogue summarization. Scientific Reports, 15(1), 1195.

Short Bio

Shlomo Berkovsky is the leader of the Interactive Medical AI research stream at Macquarie University. The stream focuses on the use of Artificial Intelligence and Machine Learning methods to develop usable patient models and personalised predictions of diagnosis and care. The stream also studies how clinicians and patients interact with health technologies and how Large Language Models can improve patient care. His other areas of expertise include user modelling, online personalisation, and behaviour change technologies.