Table 3.

Summary of core ethical principles, potential threats and possible solutions

Core ethical principles	Explanation	Potential threat	Possible solution
Autonomy	The right of the patients to make their own choices	Data breach Dynamic informed consent	Complying with the GDPR Under-representation Authorizing patients on their data Training radiology professionals about data ownership issues Making patients aware of how their data will be used
Beneficence and non-maleficence	Being impartial toward and avoiding harm to the patients	Implicit bias in the data AI algorithms that are programmed to increase the profits of their designers	Reducing the implicit biases in training data Increasing the awareness of radiology professionals and public about potential harms that can arise from AI applications
Justice	Fair distribution of medical goods and services	Different interests between developer and user Incomplete compliance with the ethical code	Factoring in unethical human interventions Training radiology professionals about ethics codes to be followed
Explicability	Transparency and accountability of the decision-making process	Ambiguity in decision-making steps Undefined accountable body	Informing the public about their rights Transparency of communication Rigorous validation of algorithms Sharing the code that is used to develop the system Training radiology professionals about how to use and produce explainable AI algorithms Increasing the awareness of public about the importance of the issue

GDPR, general data protection regulation (of the European Union).