The ethical dimension of the MELISSA research

MELISSA strives to build the first automated AI-driven treatment personalisation and optimisation platform. This activity raises serious ethical issues with regard to how much autonomy is granted to the patients and whether bias in algorithms is avoided. Although applicants ensure that the patients will be informed when interacting with both AI components and notified when AI makes recommendations, there is still a serious risk of restricting the autonomy of patients when AI is suggesting treatment.

Therefore, MELISSA has a strict ethical approach set in place to be followed rigorously. This includes the appointment of an external independent ethics advisor with competence in AI systems who reports to the European Commission.

Furthermore, an ethical framework will be established by conducting a dedicated study on the ethical impacts of MELISSA. The ethical aspects take special considerations of end-users using a trustworthy AI-driven mobile app, with a special focus on under-represented populations (minors, minorities, etc.). The study will examine and evaluate all relevant legal frameworks, literature on ethical issues related to AI-enabled health tools to ensure the performance of the MELISSA application without biases. The results of the ethical assessment will be presented in the form of recommendations for the trustworthy implementation of AI-driven diabetes management apps, addressing ethical obstacles and proposing insights to mitigate them.