The use of Artificial Intelligence (AI) tools to take notes can change the nature of interactions between patients and clinicians, sometimes missing vital information, according to the results of a new review.
The report also states that use of these tools, referred to as ambient AI scribes, may lead to deskilling and deprive clinicians of opportunities to acquire clinical reasoning skills.
While ambient AI scribes can help with administrative work, more research needs to be done to understand how these tools work in specific healthcare settings and how their design supports both patients and clinicians, experts say.
Some 40 per cent of GPs in the UK have said they used ambient scribes, which integrate speech-to-text technology with AI to assist with capturing patient notes during a consultation.
Researchers from the University of Edinburgh conducted a review of 27 articles in scientific and medical literature to examine aspects of AI scribes in terms of their adoption and the implications of scaling up their use.
One of the key findings of their analysis was that although the AI tools can support in capturing what was said by patients, they can sometimes miss out on important aspects of consultations such as facial expressions, gestures and emotional state.
AI summaries tend to prioritise clinical information, which can come at the expense of the patients' stories and experiences of their illness.
It was also found that patients may be hesitant to reveal sensitive information, such as substance abuse, domestic abuse or mental health struggles, when they know the consultation is being recorded and processed by AI.
The team also describe how manual notetaking serves important functions for clinicians in clinical reasoning and reflection on what the patient has said.
Using AI scribes could result in outsourcing mental effort – known as cognitive offloading – by clinicians. This can have benefits, such as creating the opportunity to have more meaningful conversations with patients and time to focus on complex tasks.
By contrast, these tools also can also reduce memory recall and skill development. Clinicians report instances where they do not recognise their own notes or remember the patient the next time they see them.
The researchers conclude that in designing these systems, they must preserve the patient's voice rather than overwrite it.
They also say more work needs to be done to understand the medium and long-term use of these AI tools in specific settings and emerging risks, particularly for tools used in countries with different healthcare systems to where the AI tools have been built.
The findings have been published in BMJ Digital Health and AI.
Dr Lucas Seuren, GAIL Fellow and Research Fellow, University of Edinburgh's Centre for Biomedicine, Self and Society, said: "Many clinicians are excited about ambient AI scribes, because they promise to cut down on paperwork. But the experiences of patients are poorly considered, and there are real risks that the patients' stories are lost. This can further disadvantage people who already face marginalisation in health and social care services."
