A study by INGENIO, a joint centre of the Spanish National Research Council (CSIC) and the Universitat Politècnica de València (UPV), proposes an ethical framework for incorporating artificial intelligence (AI) into clinical decision-support systems without displacing human judgement and responsibility. Published in the journal Science and Engineering Ethics, the study argues that AI-based systems should be designed as tools to support human judgement. To this end, it proposes applying a framework based on discourse ethics to incorporate ethical reflection into the decision-making process.
These computer-based tools, known as clinical decision support systems (CDSS), analyse clinical information and provide healthcare professionals with data, predictions or recommendations that can help them make decisions about a patient’s diagnosis or treatment.
Their integration into medical practice raises new ethical questions: to what extent should an automated recommendation be followed in a clinical decision? How can an outcome generated by an algorithm be justified? And who bears responsibility for the final decision? Respecting autonomy, as well as avoiding dehumanisation and the dilution of responsibility, are among the main challenges in integrating AI into clinical decision-making.
According to the study, discourse ethics is based on the idea that rules affecting people should be rationally justified through dialogue and the participation of those affected by them. Félix Lozano, a researcher at INGENIO and author of the study, explains: “Applied to the clinical setting, this approach seeks to ensure that AI-generated recommendations can be understood, questioned and assessed as part of a process of human deliberation.”
How can ethics be incorporated?
The study argues that ethical reflection should be incorporated from the design and development of these systems and maintained throughout their implementation. This process should involve not only software developers, but also healthcare professionals, ethics experts and those affected by the system’s decisions, including patients. “AI in clinical decision-making should not be assessed solely on the basis of its accuracy or predictive capabilities,” says Lozano. The researcher adds that these tools should be designed to safeguard the autonomy of both patients and healthcare professionals and enable them to understand and question the recommendations they generate.
For clinical decision-support systems to support ethical decision-making, they should function as tools that facilitate deliberation and provide relevant information for making a decision, without replacing the judgement of those involved. Their recommendations should be understandable and traceable, so that they can be assessed by healthcare professionals. Responsibility for decisions would therefore remain with healthcare professionals, while preserving patients’ informed autonomy.
To move towards this model, the study highlights the need to incorporate ethical reflection throughout the development and implementation of these technologies. According to the author, this requires interdisciplinary work bringing together fields such as moral philosophy, data engineering and public policy. This research was supported by the Regional Ministry of Culture, Education and Science of the Generalitat Valenciana (CIBEST/2024/270).
Reference: Lozano, JF. Discursive Ethics as a Normative Foundation for Integrating Ethics into AI Clinical Decision Support Systems. Sci Eng Ethics 32, 36 (2026). https://doi.org/10.1007/s11948-026-00603-1
