Study reveals how formerly incarcerated people want AI for parole navigation
31 formerly incarcerated participants reimagine parole tech as tools for strategic agency, not surveillance
A new study from Saiph Savage, Jesse Nava, Wanqing Iris Zhou, and Hwijoon Lee challenges the prevailing narrative around AI in the criminal justice system. The researchers surveyed 31 formerly incarcerated individuals about their parole experiences and what technologies they would actually find useful. The results upend assumptions that justice-impacted people want tools to dismantle the prison system entirely.
Instead, participants envisioned computational tools as instruments for navigating existing power structures with more agency. They wanted technologies that translate complex parole jargon into culturally familiar terms, help document personal growth in ways parole boards recognize, and make visible the often-unseen labor of families supporting incarcerated loved ones. The authors reframe these findings as a call for human-centered design that prioritizes strategic navigation over surveillance. The paper is set to appear in the 2026 ACM Conference on Fairness, Accountability, and Transparency (FAccT '26) and is available on arXiv.
- 31 formerly incarcerated people surveyed about parole technology needs, rejecting both surveillance tools and prison abolition assumptions
- Desired features: translating parole concepts into familiar terms, documenting transformation for boards, and surfacing family labor
- Study proposes reframing parole AI as tools for 'strategic agency' within carceral power structures, presented at FAccT '26
Why It Matters
Human-centered design can reshape parole AI from surveillance to empowerment for justice-impacted communities.