As artificial intelligence tools increasingly enter the operating theatre, questions of patient acceptance, ethical responsibility and methodological rigour become paramount. Middle East Health speaks to Sara Ben Hmido, a Dutch-Moroccan physician-researcher at Amsterdam UMC, about her team’s work examining AI in surgery from patient-centred, ethical and methodological perspectives – and why the future of surgical AI depends on enhancing human capability rather than replacing it.
Middle East Health: Could you tell us where you’re from originally, and what influenced your decision to pursue a career in medicine and surgical research?
Sara Ben Hmido: I am originally Moroccan, and my roots lie there, but I was born and raised in Amsterdam, The Netherlands. So I like to say I am a proud Dutch-Moroccan.
I have wanted to become a physician ever since I was a little girl. Helping people get better has always given me a deep sense of satisfaction and purpose.
My interest in research is closely connected to that. As a physician, you have the privilege of helping people one-to-one, which is incredibly rewarding. But as a researcher, your impact can extend far beyond the individual. You have the potential to help people one-to-unlimited, and that broader reach, the ability to contribute to knowledge and improve care on a larger scale, gives me an immense sense of motivation and fulfilment.
Middle East Health: Was there a key experience that led you to choose surgery as your area of focus? Can you tell us about it?
Sara Ben Hmido: It wasn’t necessarily one key experience that led me to choose surgery, but rather the nature and impact of the field itself. Surgery is one of the few medical specialties where, within a single procedure, you can save or significantly improve a patient’s life in one go. That immediacy and the tangible results of your work are incredibly powerful and deeply motivating to me.
Middle East Health: Can you walk us through your academic journey – where did you complete your undergraduate studies, and what steps led you to Amsterdam?
Sara Ben Hmido: I started my medical career at Amsterdam UMC, beginning my Bachelor of Science in Medicine there in 2020. Since then, it has been my main home base hospital, and it still is now in 2025. I have also completed some of my clinical rotations in other hospitals around Amsterdam, which allows me to experience different healthcare settings and patient populations.
What initially drew me to Amsterdam UMC was its personal approach to both patients and students. I really value how the institution combines cutting-edge research and innovation with genuine attention to the individual, whether that is a patient in need of care or a student developing as a future doctor. That balance between scientific excellence and human connection is something I truly admire and hope to carry forward in my own career.
Middle East Health: As a woman from Morocco pursuing advanced medical training in Europe, what unique perspectives do you bring to your research and clinical work?
Sara Ben Hmido: My Moroccan background has given me a valuable cross-cultural perspective. It allows me to approach both medicine and research with cultural awareness and resilience, valuing diverse patient experiences and striving to promote equity and understanding in every aspect of care.
Middle East Health: From recently published papers you have authored with your team, the focus is on the use of AI in the surgical field. Can you tell us in some detail about your research and the findings you have made.
Sara Ben Hmido: Our team takes a holistic approach to understanding how artificial intelligence (AI) is reshaping surgical practice. Technical innovation alone is not sufficient for meaningful progress. For AI to be responsibly integrated into surgery, it must be transparent, ethically grounded, and aligned with patient values. Through three recent studies, we have examined AI in surgery from patient-centred, ethical, and methodological perspectives to define what trustworthy and holistic surgical AI should look like.
In “Patient Perspectives on AI-Based Decision Support in Surgery” (BMJ Surgery, Interventions & Health Technologies, 2025), we examined how patients who had undergone colorectal surgery viewed intraoperative predictive machine learning (IPML) tools. Using focus groups and interviews with 19 participants across Europe, we found support for AI as an adjunct to surgical expertise, with patients describing it as a “third surgeon in the room.” Participants consistently stated that the surgeon must retain final authority. Trust in the human decision maker, clear communication about how AI influences choices, and discussion of quality of life outcomes were essential for acceptance. Participants requested accessible explanations and preoperative discussions supported by visual materials. When transparency and trust were present, patients accepted AI as beneficial; when absent, they expressed concern about loss of agency. The study demonstrated that integrating AI into surgical decision making must preserve the human connection, reinforce the surgeon’s role, and place communication with patients at the centre.
Within the ethical domain, “Ethical Pitfalls in AI-Based Predictive Models in Surgery” (World Journal of Surgery, 2025) addressed the responsibilities that accompany AI in healthcare. Predictive algorithms operate within a “stakeholder triangle” linking patients, developers, and clinicians, each with distinct obligations. Patient data form the foundation of AI, but this raises concerns about consent, privacy, and misuse, particularly if data are repurposed beyond clinical care. The study identified that biased datasets can amplify inequities, resulting in models that perform well for some populations but poorly for others. To mitigate these risks, we recommended out of distribution detection to flag unreliable predictions and the adoption of standardized frameworks such as HL7, SNOMED CT, and LOINC to improve consistency. However, ethics cannot be reduced to technical safeguards alone. Developers should disclose limitations, clinicians must remain accountable, and education in data literacy and ethics should form part of surgical training. AI should assist surgeons without replacing their judgment.
In terms of methodological rigor, “Transparency in Externally Validated Models: A systematic review of machine learning versus logistic regression for predicting colorectal anastomotic leakage” (European Journal of Surgical Oncology, 2025) evaluated the performance and reporting quality of AI models predicting colorectal anastomotic leakage (CAL), a major postoperative complication. Of 4,900 studies reviewed, only ten met strict inclusion criteria with proper external validation. Transparency, assessed using TRIPOD+AI guidelines, averaged 45 percent for logistic regression and 43 percent for machine learning models. Despite the ability of machine learning to manage complex data, its clinical reliability was limited by small, biased datasets and incomplete reporting. Logistic regression models, while simpler, proved more interpretable and consistent. The findings support developing both model types in parallel using identical datasets, ensuring external validation, and following standardized reporting frameworks before any model can safely guide surgical care.
In my view the future of AI in surgery lies not in automation but in enhancing human capability, enabling surgeons to deliver safer, fairer, and more informed care. Progress will depend on integrating technical accuracy with ethical oversight and patient trust. As AI systems become more advanced, collaboration between data scientists, clinicians, and patients will be essential to ensure these tools are developed and applied in ways that genuinely improve surgical outcomes and uphold professional integrity.
