You are here Resources > Generative AI Hub > Examples from the Disciplines

The potential and risks of GenAI applications in clinical settings

• Dr Cuisle Forde Dr Forde is an assistant lecturer in Physiotherapy and Director of the School of Medicine Online Learning Committee.

Context

In 2023/24 we added Artificial Intelligence (AI) as a topic to our module entitled “Professional issues” which is taught to Physiotherapy students. As AI was not previously taught, we grouped all senior fresh and junior sophister students together for a class on this topic. The timetable afforded us one hour of in-person teaching this year, but we will expand this for future cohorts. There were approximately 75 students in total in the class between both year groups.

What was your goal in utilising GenAI as part of the teaching process?

The aim of the session was to facilitate an open discussion around the opportunities and limitations of GenAI.
Learning outcomes were to:

• identify opportunities that GenAI tools present generally and in the clinical setting;
• discuss potential for bias, inaccuracy, misinformation and ethical issues.

As this was our first year to integrate GenAI into the module, it was not aligned with assessments and the learning outcomes were deliberately “low” on Bloom’s Taxonomy of Learning. In advance of next academic year, the curriculum will be reviewed and greater clarity will likely emerge regarding the learning outcomes related to GenAI within the module.

We added a session on GenAI as we are aware that GenAI is used in clinical practice and is highly relevant to our profession. We feel that, as educators, we need to equip our students with the best chance to enhance patient care, and AI is now part of that, for better and worse!

How did you use GenAI to enhance teaching, learning and/or assessment?

We asked students to perform a SWOT (strengths, weaknesses, opportunities, and threats) analysis before the class outlining the opportunities, risks, strengths and weaknesses of AI for the profession of Physiotherapy.

We then outlined a case study of a patient who was underweight, with a persistent cough and reported hand weakness. We asked the Kemtal GenAI platform to provide a potential physiotherapy diagnosis for this patient with the aim of:

1. identifying covert red flags (i.e. hidden signs which might suggest the presence of a serious pathology) - in this case covert red flags included weight loss and a persistent cough);
2. undertaking clinical reasoning. This is a skill required by all Physiotherapists which consists of gathering relevant information, such as signs and symptoms, and analysing that information in order to make sense of what is happening at a physiological level in a patient. Clinical reasoning helps ensure safe and effective clinical practice but it is often difficult for students to do on clinical placement. GenAI could hold potential for supporting students and clinicians with this important activity.

Moving on from the case study we discussed academic integrity, the potential for plagiarism through GenAI use and consequent impacts on the profession of Physiotherapy. We asked students to discuss potential solutions to these issues especially when in the context of take-home assignments.

We then began to move – literally! Students trialled a GenAI exercise program (see resources below) which corrects movements in real time and provides feedback on exercising to the patient as well as to the therapist who is prescribing the exercise.

What were the outcomes of using GenAI in this way?

The selected case study resulted (by design) in a contentious issue arising around the professional remit of a Physiotherapist which was subsequently debated in class. In short, the AI platform gave starkly different answers compared to those provided by the Physiotherapist. This activity showcased both the potential of GenAI to support diagnostic and clinical reasoning processes while also highlighting bias and potential to impact negatively on patient  treatment and care.

Subsequent discussions on academic integrity etc resulted in interested and wide-ranging comments and opinions!

The GenAI exercise programme was by far the most fun part of the class. Students used Vevox polling software throughout this session and when asked if they could see a use for a programme such as the GenAI exercise programme in their clinical practice they agreed that it would be very useful and that they could see themselves using something like it!

We ended by discussing if, as a profession, we should contribute to the content that GenAI uses, and, if so, how? We also questioned if it was possible for us to alter algorithms and questioned who or what was really leading to outcomes produced. That is unfortunately all we managed to cover in an hour but as part of the SWOT analysis conducted prior to class, students came up with some other ideas which we could incorporate into our curriculum. These included asking an AI bot to pretend to be a patient and then having a conversation with that patient: the transcript of this conversation could then be assessed by the academic.

There was no formal evaluation of the session, but Vevox poll contributions indicated a high level of engagement with some students reaching quite an impressive depth of thinking/consideration and others enjoying the lighter side of the topic!

What did you learn as part of this process and is there anything you would do differently?

I learnt that students themselves can, and will, tell us how to integrate GenAI into the curriculum and what they need to know if we simply ask them the correct questions. Next year I will dedicate more time to this topic and we will be able to teach one year group at a time (as it won’t be the first year of this topic being introduced). Smaller numbers will also help facilitate better discussions.