Research published in the Journal of Clinical Medicine validates Supportiv’s digital peer support solution as a promising intervention for early detection of suicidal ideation.

The hybrid AI model + human triage solution represents a powerful tool to address suicidal ideation, as the tool’s AI rapidly detects suicidal risk, while the human moderator immediately responds, in the midst of an ongoing digital peer support experience, which delivers meaningful emotional improvement.

Read the full study text here.

This study examined the effectiveness of Supportiv’s hybrid solution combining a proprietary AI suicidal ideation (SI) detection model with real-time human moderation within 449,946 user sessions from January–December 2024.

Key paper findings based on passive and active SI instances identified using a hybrid AI and human moderator solution with post-hoc LLM verification:

  • Active and passive SI occurred in 3.19% of live chats
  • The proprietary AI model identified SI faster than humans in 77.52% of passive and 81.26% of active cases, with 90.3% agreement
  • Within 71.3 seconds after AI identification, moderators had enacted crisis question protocols to determine active vs. passive SI and referred 5,472 active SI users (1.21%) to crisis care
  • Users participating in live chats with SI revelations saw average reductions of 29.3% in depression, 26.8% in loneliness, 25.3% in despair, and 22.3% in helplessness, with optimism increasing on average 40.4%

The paper was featured as part of a rigorously-reviewed special issue showcasing industry-leading research on Machine Learning and Artificial Intelligence in Clinical Medicine.

Read more published research on Supportiv’s digital peer support solution.