AI Ethics

Clinical AI ethics applies core ethical principles -- such as autonomy, beneficence, non-maleficence and justice -- to the design and use of artificial intelligence tools in patient care and clinical or population decision-making.

Examples of adaptations of clinical ethics principles include:

• Autonomy
  informing patients and providing opportunity to refuse consent when AI significantly influences communication about or conduct of diagnosis, prognosis, or therapy.

• Beneficence
  designing and implementing AI tools in ways that improve the health processes or outcomes that patients, and their caregivers, most care about.

• Non-maleficence
  desgining, implementing and montitoring AI tools to anticipate, identify and prevent harm, avoid bias, and minimize error (hallucinations).

• Justice
  consider and avoid AI alignment mismatches that could entrench or increase racial, demographic or geographic inequalities.