Human-Led Artificial Intelligence (AI) for Corrections: Using Hapi, A Generative AI Digital Companion to Support Discharge Planning, Community Reintegration, and Mental Health (PID040)

2pm – 2.30pm WEST, 23 April 2026 ‐ 30 mins

Parallel Workshops

Correctional leaders are grappling with increasing pre-trial populations, chronic staff shortages, high rates of burnout, and overwhelming service requirements needed to support their complex justice-involved populations. This session introduces Hapi—a human-led, dual-system artificial intelligence powered digital companion—as a blueprint for scalable solutions across four correctional domains: Discharge Planning, Crisis Intervention,  Mental Health, and Community Supervision    

Hapi is purpose-built to enhance human staffing, providing 24/7 trauma-informed support to residents and community corrections clients via secure closed-network tablets or SMS. Utilizing AI for personalized responses and ethical triage, Hapi initiates comprehensive discharge planning months before release, assisting residents with complex resource navigation for housing, ID, employment, and health services. Crucially, Hapi retains conversational memory and confidentiality across environments enabling the resident's discharge plan to seamlessly transfer from custody to community, establishing a personalized continuum of care between environments.

Hapi simultaneously mitigates staff burnout by offloading time consuming tasks of responding to routine queries, providing quick access to internal protocols, and when deployed for staff wellness, Hapi offers responders a confidential, 24/7 mental health support tool that is accessible when they need it - while on the job or at home. Human-in-the-loop guardrails ensure ethical boundaries, and the system’s triage capabilities trigger immediate human alerts for high-risk crises like self-harm or overdose helping to mitigate the tragic number of people dying in custody. The session will provide a decision framework, live scenario walkthroughs, and an implementation checklist, demonstrating the feasibility of this multilingual, low-bandwidth technology to deliver measurable improvements for both staff and residents.