Artificial Intelligence Must Operate Ethically in Health Care and Not Be Prone to Racist or Sexist Biases

This is the authors' abstract. We don't add key points for this paper.

Artificial intelligence (AI) promises to transform modern health care on multiple fronts, which may continue the long history of health care improvement enabled by the advancement of technology. 1 However, the complexity of AI systems poses questions regarding their reliability and safety in the health care context, particularly in the light of recent evidence of biases in the operation of AI systems.2 Additional concerns include uncertainty around the strengths and limitations of AI systems, and the lack of clear strategies for the safe integration of AI with existing health care technology, presenting clinicians and health care leaders with difficult and potentially very expensive choices. In the following, we outline the underlying nature of AI systems, their limitations, and the risks that their use may pose in health care and anesthesia.