Using AI to Detect and Prevent Suicide
Author Information
Author(s): Levkovich Inbar, Omar Mahmud
Primary Institution: Tel-Hai Academic College
Hypothesis
Can large language models improve the detection, prevention, and risk assessment of suicide?
Conclusion
Large language models show significant potential in identifying and detecting suicidal behaviors, often outperforming mental health professionals.
Supporting Evidence
- Most studies demonstrated that large language models are highly efficient in detecting suicidal ideation.
- LLMs often outperform traditional mental health professionals in early detection and prediction capabilities.
- Ethical concerns regarding the use of AI in mental health need to be addressed.
Takeaway
This study looks at how computers can help find people who might want to hurt themselves and how to stop that from happening.
Methodology
The review systematically searched seven databases for studies published from January 1, 2018, to April 2024, focusing on the use of large language models in suicide prevention.
Potential Biases
Some studies exhibited cultural and gender biases, affecting the applicability of the results.
Limitations
Many studies relied on synthetic data and had small sample sizes, which may limit the generalizability of the findings.
Participant Demographics
The studies included various demographics, primarily focusing on social media users and clinical populations.
Statistical Information
P-Value
p<0.001
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website