Artificial intelligence as a tool for cognitive impairment screening: Patient perspectives about benefits and limitations

Abstract Introduction This paper presents findings from a qualitative study exploring the perspectives of primary care patients on the use of artificial intelligence (AI) to synthesize data from the electronic medical record (EMR) and audio recordings of patient speech as a way to detect cognitive impairment. Methods We conducted qualitative interviews with a racially and educationally diverse sample of primary care patients ( N = 31) from a large healthcare system in New York City, New York. Interviews focused on the use of AI to synthesize data from the EMR and audio recordings of patient speech. Data were analyzed using content analysis. Results Participants reported that AI technology could improve diagnostic accuracy and avoid human errors in the evaluation and diagnosis of cognitive impairment, while also improving clinical efficiency for clinicians. They also expressed concerns about the use of AI tools for screening, such as risk of misdiagnosis, loss of confidentiality, and psychological distress from learning of detected cognitive impairment. Conclusion The perceived benefits reported by participants regarding the use of AI screening tools suggest that some patients might support the provision of AI screening in clinical settings. However, participants’ concerns about the potential harms of AI screening tools underscore the need for clear and transparent communication about how results are generated and interpreted. Clinicians need training and other forms of structural support to improve their literacy in AI‐driven tools and their ability to integrate adequate approaches to information sharing into clinical practice.