Table 2.

Illustration of the framework for chatbot risks to the principles of medical ethics.

Risks	Ethical principles
	Beneficence	Nonmaleficence	Autonomy	Justice
Errors	—a	The chatbot makes the wrong recommendation to patients based on a bug in the system.	—	—
Discrimination	—	The chatbot has a bias that causes it not to understand requests based on women’s health.	—	The chatbot provides more appropriate recommendations for men than for women.
Stereotyping	The chatbot responds to the patient in derogatory terms.	The chatbot’s recommendations based on stereotypes lead to harm the patient.	—	The chatbot gives unfair and derogatory responses to patients.
Exclusion	—	The chatbot excludes certain users because of language and literary skills, withholding medical support.	—	The chatbot excludes certain patients and no alternative is provided.
Stigma	—	Use of the chatbot is not anonymous and leads to stigmatization of certain patients.	—	—
Lack of privacy	—	—	There are data leaks from the chatbot system leading to a breach of confidentiality.	—
Poor data governance	—	—	Patients do not consent to have their data collected by the chatbot and mechanisms for data governance are not clear.	—
Overconfidence and trust decay	—	The chatbot harms the relationship between the patient and their physician by providing contradictory recommendations.	—	—
Technological solutionism	A chatbot is not the best option for providing medical recommendations to certain patients.	—	—	—

^aNot applicable.