Artificial intelligence must help patients become more autonomous by giving them access to clear, reliable and tailored information. This allows them to better understand their situation, take an active role in their health choices and decide in collaboration with professionals.
AI must strengthen the quality of care and must never reduce healthcare professionals to the role of mere technical operators.
Algorithmic biases and the potential inequalities they generate must be actively identified and corrected to ensure fair access.
Data confidentiality and consent must be strictly respected at every stage (collection, use and restitution of data).
BST Thinkers advocates for innovation that puts humans back at the center, while respecting the fundamental values of medicine and research.
The Expansion and Ethical Challenges of AI
Complete White Paper: Argument and Recommendations
Artificial intelligence opens vast fields of application in medicine: advanced imaging, rapid diagnostics, and personalized risk prediction. These innovations raise questions about the quality of care and the human relationship, which must remain at the heart of medicine. AI should support medical decision-making but never replace the practitioner’s judgment, who remains medically responsible.
Training healthcare professionals—both during their studies and throughout their careers—is essential to strengthen their skills and their ability to use AI in a critical and informed way.
Bioethical Principles Adapted to AI
Patient autonomy: Ensure accessible information and free, informed consent.
Beneficence and non-maleficence: Promote well-being and avoid harm arising from automated decisions.
Justice and equity: Guarantee equal access to the benefits of AI for all.
Confidentiality: Strictly protect data from misuse.
Dilemmas: Decision Support or Substitution?
The central issue lies in preserving clinical judgment and expertise in reasoning and decision-making.
In practical know-how:
AI should support, not replace, the practitioner’s judgment. The physician remains responsible for medical decisions. Training must therefore reinforce healthcare professionals’ technical competence and their ability to use AI knowledgeably and critically.
In relational know-how:
AI must be used without ever replacing the practitioner’s discernment. A physician must combine technical expertise with relational skills. Training should thus also develop empathy, communication, and the ability to build continuous dialogue with patients. This approach fosters mutual understanding and enables patients to anticipate and actively engage in decisions about their health.
AI Bias: Continuous Monitoring and Correction
To ensure ethical and reliable AI use, it is essential to understand and detect biases. Training data can inadvertently exclude population groups—based on age, gender, origin, or socioeconomic status—leading to diagnostic errors. Such biases can also reinforce health inequalities, particularly when certain populations have reduced access to care, technology, or medical literacy.
Several levers can limit these risks:
Diversify data sources.
Rigorously validate algorithms.
Implement continuous oversight and regular evaluations.
These measures are necessary to ensure fairness and uphold medical justice for all.
Responsibility and Legal Framework
Roles must be clearly defined: the physician always retains ultimate validation over decisions assisted by AI. A clear regulatory framework, including alignment with the European AI Act, is essential to protect both patients and practitioners.
Ethical Data Management
Informed consent and transparency in data use—for care, research, or industrial applications—are indispensable. Any secondary data use must be clearly communicated and secured to preserve patient trust.
Ethical responsibility is collective: designers, manufacturers, and users must all uphold the same standards of rigor and respect for rights.
Data management raises particularly complex questions in emerging fields such as bioprinting. In this domain, several players are involved—the algorithm designer, the device manufacturer, and the practitioner who implants or uses the final product.
Across invention, production, and application, responsibility becomes diluted, especially when biological and digital data intersect. Who is accountable in case of an error, misuse, or unauthorized data reuse?
Currently, no universally accepted answer exists. This uncertainty highlights the urgent need for an ethical and legal framework capable of defining collective responsibility while ensuring transparency, traceability, and informed patient consent at every stage of the innovation chain.
BST Thinkers Recommendations
Raise awareness and train professionals in both ethical and critical uses of AI, alongside technical learning.
Ensure data representativeness and fairness from collection to processing, guaranteeing balanced access to digital innovations.
Maintain human medical judgment as the ultimate decision criterion, supported by explicit, transparent reasoning that can be communicated to patients.
Comply with European legal frameworks and international guidelines on AI and data governance.
Develop “soft skills” such as trust, responsibility, empathy, and communication to preserve the human dimension of medical practice in the digital age.
Strengthen professional confidence and interpersonal skills through dedicated training.
The future of AI in medicine will depend on responsibility and humanity. Innovation must enhance care and trust, serving patients, professionals, and the common good—without sacrificing core human values.