Harnessing artificial intelligence to capture online mental health stigma: A scoping review

Mental health stigma remains a major barrier to care and is increasingly expressed online. Artificial intelligence (AI) may offer scalable ways to detect and analyze stigma, providing insights that could inform stigma-reduction strategies.

This scoping review aimed to map the existing literature on the use of AI to capture online mental health stigma, highlighting key approaches, gaps, and directions for future research and interventions.

Following PRISMA-ScR guidelines, we searched Web of Science, PubMed, PsycInfo, and IEEE Xplore for English-language peer-reviewed studies published between January 2015 and June 2025. Studies were included if they applied AI to analyze online public discourse related to mental health stigma.

33 studies met inclusion criteria. AI models processed large volumes of online text to detect stigmatizing content and identify themes in mental health discourse. The most frequently used AI techniques were sentiment analysis and topic modeling. Few studies used context-sensitive AI models capable of capturing subtle cues such as sarcasm that signal stigma. Research focused mainly on English-language content and common conditions such as depression and substance use disorders.

AI holds promise for identifying and analyzing online mental health stigma at scale. However, current approaches often conflate stigma with superficial tone, lack validation across diverse settings, and underutilize newer AI models that are capable of capturing the nuances of stigmatizing language. Future research should address these gaps to guide the development of more effective anti-stigma interventions.