Mental health disorders significantly impact quality of life and are a leading cause of disability, often showing up in the way people use language on social media. The past few years, especially since COVID-19, have seen a rise in developing methods to detect these signs in social media language, which might soon help predict mental health issues reliably. This could be crucial for early mental health interventions. This study looks into the latest artificial intelligence techniques, like natural language processing, which analyze vast amounts of text to predict mental health conditions. However, research in this field is limited by a lack of quality data and the need for strong ethical and privacy measures to protect those with mental health issues.
The research reviewed articles from 1999 to 2024, focusing on data sets, prediction methods, long-term studies, and ethical considerations. It found a need for more detailed data, particularly from people with severe conditions like schizophrenia. Encouraging users to share their social media data could address data shortages and privacy concerns. Besides text, other methods involving voice, images, and videos might offer better predictions and gain more public trust. More robust social media data and careful attention to privacy are necessary to advance these methods. The study suggests a social media feature allowing users to contribute their data for research and calls for further investigation into how social media affects depression and anxiety.
Access the article here.