Cancer clinicians welcome AI that supports human expertise

Artificial intelligence is often presented as a threat to professional work, raising fears about deskilling, job displacement and the loss of human judgement. But new research involving healthcare professionals using AI in cancer treatment suggests a more nuanced story.

The study, published in Sociology of Health & Illness, found radiotherapy professionals were overwhelmingly positive about AI tools used to help contour organs-at-risk during cancer treatment planning.

In radiotherapy, clinicians must identify and protect healthy organs close to a tumour so radiation can be targeted safely. This process, known as contouring, can involve drawing the boundaries of organs-at-risk across many scans. The paper describes this as highly skilled work, but also repetitive, time-consuming and, in many cases, tedious.

Researchers assessed the lived experiences of 32 healthcare professionals across five regional cancer treatment centres in England. Participants included clinical scientists, radiographers, dosimetrists and oncologists involved in radiotherapy planning.

Interviewees did not believe AI had taken over their expertise. Instead, they saw it as taking on the first draft of a burdensome task. Professionals still reviewed the contours, corrected errors and retained responsibility for the final treatment plan.

One clinical scientist told the researchers that before AI, drawing around organs manually "would take between half an hour and two hours per patient". The AI tools could produce a first draft in "five or ten minutes", removing a task that "quite highly trained, quite in demand people had to do".

The researchers describe this as a "partial discard" of work. AI did not remove the task from professionals entirely. It removed the most routine stage, while preserving professional oversight, accountability and authority.

This helps explain why the technology was accepted so positively. Interviewees repeatedly stressed that the AI could make basic mistakes, especially when anatomy or local clinical practice differed from what the system expected. Rather than undermining professional expertise, those mistakes reinforced the need for human judgement.

Interviewees said AI allowed them to spend more time on higher-value activities such as treatment planning, service improvement, research and more personalised patient care.

One oncologist put it sharply: "People worry about it replacing clinicians. I don't think our expertise is drawing around lumps and bumps."

The findings suggest debates about AI and work need to move beyond simple claims that technology will either replace professionals or liberate them. The impact depends on the task, the workflow, the professional group and whether the technology helps people do the work they value most.

Read the full study, Implementing AI in Radiotherapy: Insights From the Healthcare Professions, by Dr Juan Baeza, Professor Jon Hindmarsh, Dr Angie Kehagia and Dr Anna Barnes.