Looking forward, I anticipate AI playing a role in almost all medical procedures in the foreseeable future. This forecast is not solely based on observations but predominantly on the numerous advantages that AI is set to bring to healthcare services, as supported by various studies.
Given this trajectory, defining the standard of care in AI-based medicine is long overdue. Establishing these guidelines is crucial to maximise the benefits of AI, ensure patient safety and prevent potential instances of medical liability. At a later stage, it is crucial to clarify the assignment of legal liabilities, not only to ensure that patients can easily receive compensation but also to enable the various stakeholders (AI developers and manufacturers, hospitals, doctors) to fully understand their respective roles.
It is vital for professional bodies and scientific associations to establish operational rules for using AI, considering the specific technical aspects within each medical speciality. This responsibility primarily rests with healthcare professionals.
However, it has long been acknowledged that the existing legal framework significantly shapes the standard of care, particularly concerning issues like preserving medical confidentiality, obtaining informed consent, and organising collaborative medical practices. Navigating this area presents a challenge: legal professionals struggle due to the novelty and disruptive nature of AI technology, alongside insufficient current regulations. Legal advancements often lag behind technological progress. The case law is still developing, and prevailing opinions in doctrine frequently lack substantial foundations.
The questions to solve are numerous and complex: Should doctors disclose to patients that AI-supported decision systems were used in diagnoses? What additional information should patients be given? How should doctors proceed when they lack the necessary information due to AI’s lack of transparency and/or explainability? Are doctors obligated to adhere to AI recommendations in diagnoses or treatments? When is it appropriate for doctors to disagree with AI recommendations? Does the use of AI undermine the trust between doctor and patient? Who is to ‘blame’ when something goes wrong: the doctor, the hospital, the AI developer?
The questions are numerous, yet comprehensive and legally sound answers are limited. This legal uncertainty poses an imminent risk to healthcare professionals, a pattern often observed during the initial stages of adopting new technology. However, it is crucial not to let this ambiguity hinder or discourage the advancement of AI in medicine. While doctors are irreplaceable, those who disregard AI may find themselves easily replaced.