Preparing for a digital future
Sarah Davies, Radiography Lecturer and Doctoral Researcher in Medical Image Perception, Institute of Health, University of Cumbria
PhD title: Understanding human-AI interaction in medical imaging: Moving towards ‘appropriate reliance’
I was recently fortunate enough to be awarded a fully funded place at the Northern Futures Digital Health and Care Summer school at Newcastle University. This three-day event brought together PhD students and early career researchers from across the country to explore how digital technologies are transforming healthcare. The timing of the event was particularly significant following the publication of the government’s 10-year ‘Fit for the Future’ plan in July, which emphasised the shift from analogue to digital as a way of addressing some of the major challenges facing the NHS. Of particular interest to me was the goal of AI becoming ‘every nurse’s and doctor’s trusted assistant’. While this might sound straightforward, establishing an appropriate level of trust in AI is often challenging, which is why continued research into human-AI interaction is so important.
The theme of trust in AI came up repeatedly throughout the summer school and it connected directly with my own research. I had an opportunity to present my work on the role of confirmation bias during an AI-assisted medical image interpretation task, the main aim of which is to investigate whether confirmatory AI output influences user perceptions of system reliability and subsequent reliance on AI. Attendees were particularly interested in the use of eye-tracking technology to highlight changes in visual search behaviour when interacting with AI. For example, the presence of an AI prompt may lead the observer to neglect other regions of the image. This may be beneficial if the prompt has correctly identified a suspicious region but detrimental if AI misses an important finding. Studying these human factors is essential if AI integration in clinical decision-making is to be both safe and effective.
Other discussions focused on the importance of ensuring that high-quality datasets are used in the training, testing and validation of AI models. Poor-quality data may not represent the target population, meaning the AI system, like a human, may also exhibit bias. Professor Amanda Taylor-Beswick delivered a session around the use of AI in social work, highlighting the significance of safe data storage and processing as well as emphasising the need for human oversight to ensure data accuracy. Within the social work setting, AI is primarily used as a means of increasing efficiency, but if we use these models without really understanding how our data is handled, or if we assume accuracy in automated data transcription, are we potentially in breach of GDPR?
Overall, attendance at the summer school was an inspiring and thought-provoking experience. Whilst we were all in agreement that a digital future is inevitable and, indeed, necessary, it is not without risk. However, events such as this, which bring healthcare professionals and researchers together to share ideas and innovations pave the way towards a digital future that is safe, ethical and effective.
Want to find out more about our Postgraduate Community?
A postgraduate research degree at the University of Cumbria offers the perfect opportunity to follow your passion, enhance your skills and influence the world around you. As the blogs here show, we are home to a rich community of people from all kinds of backgrounds and career points. Rated as the best University in the UK for overall PhD experience in 2025, Cumbria offers a welcoming, supportive and challenging environment to utilise your enthusiasm and knowledge and turn it into insightful and impactful work. If you are interested in undertaking a research degree with us, then please contact Professor Tom Grimwood via Tom.Grimwood@cumbria.ac.uk for a conversation.
.jpg)
.jpg)