Published online Oct 19, 2022. doi: 10.5498/wjp.v12.i10.1287
Peer-review started: July 11, 2022
First decision: August 1, 2022
Revised: August 9, 2022
Accepted: September 22, 2022
Article in press: September 22, 2022
Published online: October 19, 2022
Processing time: 98 Days and 11.6 Hours
Artificial intelligence-based technologies are gradually being applied to psych-iatric research and practice. This paper reviews the primary literature concerning artificial intelligence-assisted psychosis risk screening in adolescents. In terms of the practice of psychosis risk screening, the application of two artificial intelligence-assisted screening methods, chatbot and large-scale social media data analysis, is summarized in detail. Regarding the challenges of psychiatric risk screening, ethical issues constitute the first challenge of psychiatric risk screening through artificial intelligence, which must comply with the four biomedical ethical principles of respect for autonomy, nonmaleficence, beneficence and impartiality such that the development of artificial intelligence can meet the moral and ethical requirements of human beings. By reviewing the pertinent literature concerning current artificial intelligence-assisted adolescent psychosis risk screens, we propose that assuming they meet ethical requirements, there are three directions worth considering in the future development of artificial intelligence-assisted psychosis risk screening in adolescents as follows: nonperceptual real-time artificial intelligence-assisted screening, further reducing the cost of artificial intelligence-assisted screening, and improving the ease of use of artificial intelligence-assisted screening techniques and tools.
Core Tip: Artificial intelligence-assisted psychosis risk screening must be emphasized and applied in adolescents. This review summarizes the application of two artificial intelligence-assisted screening methods (chatbot and large-scale social media data analysis), and proposes that the first challenge in applying artificial intelligence to psychosis risk screening concerns ethical issues. The methods must follow four biomedical ethics principles, i.e., respect for autonomy, nonmaleficence, beneficence, and justice. Three directions should be considered in the future: nonperceptual real-time artificial intelligence-assisted screening, further reducing the cost of artificial intelligence-assisted screening, and improving the ease of use of artificial intelligence-assisted screening techniques and tools.