You are preparing for your ST6 interview. The question about AI is coming — it comes up in almost every panel now. The problem is not that trainees don’t know anything about AI. It’s that most are not sure which parts of what they know actually matter.
This post is the answer to that question.
What AI in surgical training already does
Before deciding what to prioritise, it helps to know where AI has already arrived in training environments. A 2022 scoping review of 49 peer-reviewed studies on AI in surgical education found that the primary applications were assessment of operative competency, personalisation of training, and feedback in surgical simulation — with many studies reporting high accuracy in objectively characterising surgical skill from video or instrument kinematics data (Kirubarajan et al., 2022).
That assessment function has been tested in a real trial. A randomised controlled trial published in JAMA Network Open compared AI tutoring with expert instructor feedback during surgical simulation training in medical students (Fazlollahi et al., 2022). The AI tutoring system improved practice Expertise Scores by 0.66 points compared with the expert instruction group (P<0.001), with equivalent cognitive load and emotional responses between groups. The AI outperformed the human instructor on the measured outcome.
AI in surgical training is not a future development. It is current. Its most active application right now is skills assessment and simulation feedback — not clinical decision support.
The skill that actually matters
The instinct is to focus on tools. Which AI is in your department. What the planning software does. Whether the PACS system flags fractures automatically. These are not wrong to know — but they are not what makes you clinically useful when a new tool arrives.
The skill is appraisal. When an algorithm appears in your clinical environment — fracture detection on call, outcome prediction before arthroplasty, preoperative planning software — the trainee who adds value is the one who knows which questions to ask. What was it trained on? Was it externally validated on a population comparable to yours? What was the comparator? What was the risk of bias in the underlying studies?
That framework does not become obsolete when the next tool launches. It applies to every AI claim you will encounter across your career. Knowing the name of five products does not.
Portfolio
You are already encountering AI in practice. The CT-based preoperative plan on a robotic list. The algorithm flag on the emergency PACS. The automated scoring embedded in simulation software. Most trainees move past these without logging them.
Logging them is legitimate portfolio material. You do not need to have conducted an appraisal yourself. A reflection that describes encountering an AI fracture detection tool on call, asking what it had been trained on, and not receiving a clear answer is a meaningful reflection. It demonstrates critical engagement rather than passive use — which is exactly what the curriculum asks for.
If you want stronger portfolio material: find and read the published evidence for one tool you have encountered in practice and write a structured critical appraisal. A learning log entry, not a publication. It demonstrates the same competency an examiner would be looking for, and almost no trainee does it.
The interview answer
A strong answer to “what’s your view on AI in orthopaedic surgery?” has three components. It names something specific. It states what the evidence shows. And it distinguishes what has been demonstrated from what is anticipated.
For example: AI fracture detection performs comparably to experienced clinicians on external validation, but automation bias — the tendency to defer to the algorithm when clinical examination says otherwise — is a documented failure mode. Robotic systems improve component positioning accuracy consistently; whether that accuracy produces better patient outcomes compared to conventional surgery has not yet been demonstrated in randomised trials.
That is a grounded, specific, one-minute answer. It cites real evidence. It names real limitations. It will stand out.
What not to say: that AI has “enormous potential” or will “transform the specialty.” Every panel has heard both. Neither is specific enough to evaluate, and both signal that you have read about AI without engaging with what the evidence actually shows.
The question is coming. The answer that impresses is not the one that shows you have heard the most. It is the one that shows you know how to think.
For the evaluation framework mentioned above, see What do we actually mean when we say AI in orthopaedics? For specific examples of automation bias and robotic TKA evidence, see The 3am fracture and Mako, ROSA, and the rest.
References
- Kirubarajan A, Young D, Khan S, et al. Artificial Intelligence and Surgical Education: A Systematic Scoping Review of Interventions. J Surg Educ. 2022;79(2):500–515. https://doi.org/10.1016/j.jsurg.2021.09.012
- Fazlollahi AM, Bakhaidar M, Alsayegh A, et al. Effect of Artificial Intelligence Tutoring vs Expert Instruction on Learning Simulated Surgical Skills Among Medical Students: A Randomized Clinical Trial. JAMA Netw Open. 2022;5(2):e2149008. https://doi.org/10.1001/jamanetworkopen.2021.49008