A family-supported educational access and reentry preparation platform for correctional environments. This document explains what is buildable now, what requires institutional partnerships, and what a realistic pilot pathway looks like.
A structured platform where families can fund and manage safe, monitored AI-assisted educational and reentry support for their incarcerated loved ones.
Families use a mobile or web app to create a profile for their incarcerated loved one, choose a support plan, fund commissary and book deposits, and generate session access codes for Beacon — the platform's purpose-limited educational AI.
After every session, they receive a human-readable summary covering topics discussed, safety flag status, progress updates, and a recommended next step.
Using an approved facility terminal, tablet, or kiosk, the incarcerated individual enters their ID and a family-generated access code. A monitored, purpose-limited educational session begins — focused exclusively on approved categories including GED preparation, financial literacy, parenting communication, wellness, job readiness, and reentry planning.
Every session is logged. Transcripts are available to authorized facility staff. No open internet. No anonymous messaging.
* RAND Corporation, BJS, Vera Institute. Cited for context — not claimed as platform outcomes.
Families need a structured way to support both immediate daily needs and long-term growth — in one unified platform.
The following features are achievable as software-layer deliverables — no prison system integration required. These form the foundation of a demonstration prototype and investor/partner conversations.
Moving from prototype to deployed product requires formal relationships with correctional institutions, approved technology vendors, and payment infrastructure providers.
The correctional technology space is dominated by a small number of established vendors with existing contracts across most U.S. facilities. Second Chance AI would need to either partner with these vendors, position as a software layer on top of their infrastructure, or pursue direct facility relationships in jurisdictions where alternatives are permitted.
One of the most emotionally significant and practically important features of the platform. Families need a way to support immediate daily needs — not just invest in long-term preparation.
Beacon is purpose-built for the correctional educational context. Every design decision is intended to maximize institutional trust and safety while providing genuine educational value.
All Beacon responses include appropriate disclaimers for legal and medical content. Beacon is not a licensed attorney, financial advisor, or medical professional. Every session begins with a monitoring notice. The system is designed to supplement — not replace — qualified human guidance. Hallucination risk is mitigated through category restriction, response calibration, and mandatory disclaimers.
The following areas require careful legal research before any pilot deployment. This is not a legal opinion — it is a research agenda. Qualified correctional law counsel should be engaged before pilot conversations.
Nothing in this section constitutes legal advice. The areas below are research topics identified as relevant — not legal determinations. All claims should be verified with qualified legal counsel before pilot deployment. Legal requirements vary significantly by jurisdiction, facility type (state prison vs. county jail), and existing vendor contracts.
Research the federal legal framework governing monitoring of inmate communications, including Title III of the Omnibus Crime Control Act, Prison Litigation Reform Act implications, and relevant case law on the extent of First Amendment protections in correctional settings (see Turner v. Safley, 1987). Key question: what disclosures are legally required at session start, and what logging/retention obligations apply?
Federal and state correctional facilities have widely varying policies on internet access for incarcerated individuals. Research current BOP (Bureau of Prisons) policies, state-level policies, and the legal basis for restrictions. Understand whether a closed, non-internet AI system faces different scrutiny than general internet access — it likely does, but this needs verification.
Research what privacy protections, if any, apply to incarcerated individuals in educational communications contexts. Understand obligations under FERPA (if educational records are created), state privacy laws, and data retention/destruction requirements. Consider whether family members' data (on the family-facing app) triggers different privacy obligations (CCPA in California, etc.).
Real commissary deposits would likely require money transmission licensing in applicable states, or a partnership with an already-licensed entity. Research FinCEN requirements, state money transmitter licensing, bank partnership structures, and whether the "agent of a payee" exemption or similar exemptions apply. AML/KYC compliance is also required for any flow-of-funds model.
Research emerging AI liability frameworks, including FTC guidance on AI deception, state consumer protection requirements, and correctional education program standards. Understand what disclaimers are legally required for AI-generated educational content, and whether the platform could face liability for AI hallucination in the legal literacy or wellness categories specifically.
The FCC has issued significant rulings on prison phone costs and oversight (see In the Matter of Rates for Interstate Inmate Calling Services). Research whether Beacon sessions could be classified as a communication service subject to FCC jurisdiction, or whether the educational nature of the platform places it outside FCC's prison communications framework.
The Second Chance Pell Experiment (expanded under the FAFSA Simplification Act) restored federal Pell Grant eligibility for incarcerated individuals. Research whether Beacon-facilitated coursework could qualify under this framework, what accreditation requirements apply, and whether partnership with an accredited educational institution is required to access this funding pathway.
California is the most logical initial pilot state given its scale, existing rehabilitation programming, and relative openness to reentry innovation. However, CDCR deployment would require significant institutional alignment.
County jails in California operate under county sheriff departments and may have faster, more accessible procurement processes than CDCR state prisons. Consider approaching Los Angeles County (LASD), San Francisco County, or Alameda County as early pilot candidates.
Key advantage: County jails often serve a higher proportion of pre-trial detainees and shorter-sentence individuals — exactly the population where reentry preparation timing matters most.
Nonprofit and community-based reentry organizations can serve as credibility partners, referral sources, and eventual program integrators. These organizations often have existing facility relationships and can help navigate institutional procurement.
The platform is built on a set of non-negotiable ethical commitments. These principles are not aspirational — they are baked into every design and product decision.
An honest, phased path from prototype to measured pilot. Each phase has clear deliverables and decision points.
A layered architecture designed for institutional trust — every component is auditable, every data flow is monitored, and every integration point is clearly delineated.
A diversified model that avoids the extractive pricing common in the correctional technology industry.
Honest risk assessment is a prerequisite for institutional trust. Every serious risk has been identified and a mitigation path defined.
A working checklist for advisor review and research prioritization. Click items to mark them complete.
Concrete, sequenced actions for the next 90 days.