A Governance Summary for Funders and Institutional Partners
Justice Education Society | June 2026
Justice Education Society (JES) is actively integrating artificial intelligence tools into our operations. This document summarizes the governance framework we have built to ensure that AI adoption at JES is responsible, transparent, and consistent with our obligations to the populations we serve, our funders, and our partners.
We share this summary because we believe funders have a legitimate interest in how their grantees manage emerging technology risks, particularly where programs serve vulnerable populations. We also believe that being explicit about our governance approach builds the trust required for effective long-term partnership.
About JES and Our AI Context
JES delivers public legal education and access-to-justice programs in British Columbia and across 9 active country programs internationally, with offices in Vancouver, Guatemala, Honduras, Guyana, and Saint Lucia. Our programs are funded by federal and provincial governments, gaming and lottery authorities, municipal governments, and through program fees and fundraising. We serve vulnerable populations including individuals involved in legal proceedings, survivors of violence, youth, and marginalized communities.
We operate Ask AI, a legal information tool built on verified content and designed to help people access legal information before speaking with a professional. We also use general-purpose AI tools (Claude and Google AI) to support internal operations including research, drafting, translation, and reporting.
Our programs operate across multiple legal and cultural jurisdictions. This means our data governance approach must meet not only Canadian standards but also the terms of our contribution agreements and the expectations of host country partners and communities.
Our AI Governance Framework at a Glance
| Governance Area | What JES Has in Place | Why It Matters |
|---|---|---|
| Data classification | All JES data is classified into three tiers. Tier 1 (prohibited) covers personally identifiable data involving clients, minors, and vulnerable populations — no AI tool may be used with this data under any circumstances. Tier 2 (restricted) requires use of privacy-protected sessions only. Tier 3 (permitted) covers aggregate and publicly available information. | Prevents inadvertent exposure of participant data through AI platforms, consistent with PIPA (BC) obligations and contribution agreement requirements. |
| Vulnerable population protections | Data involving minors, survivors of violence, and individuals in legal proceedings is classified Tier 1 regardless of form. Re-identification risk in small program cohorts is explicitly accounted for. | Programs serving vulnerable populations carry heightened obligations. Our classification rules go beyond minimum legal requirements. |
| International program data | Program activity data from specific countries is assessed for host country sensitivity before any AI use. Information shared only between JES, partners, and funders is treated as confidential regardless of whether it contains personal identifiers. | Protects partner relationships and community trust. Ensures country-level information does not enter AI platforms where it could be exposed or used for training. |
| Human oversight requirement | All AI-assisted outputs submitted, published, shared externally, or used in program contexts must be reviewed by a qualified staff member. The use of AI does not reduce professional accountability. | Ensures quality and accuracy in all program-related work. Prevents unreviewed AI output from reaching funders, clients, or program participants. |
| Candidate and applicant data | AI tools may not be used to screen, rank, or shortlist job candidates. Applicant personal information (names, addresses, contact details) must not be uploaded to AI platforms without explicit policy authorization. | Addresses bias risks and privacy obligations relating to personal information provided in confidence during hiring processes. |
| Consultant and third-party quality | JES expects consultant deliverables to reflect genuine expertise. AI-assisted work that substitutes for professional judgment is not acceptable. Staff are directed to flag quality concerns for any deliverable that appears to have been produced primarily through AI without expert input. | Protects the integrity of program delivery and the reliability of expert advice on which programs depend, particularly in legal, financial, and specialized technical contexts. |
| Cloud storage and platform integration | Direct AI integration with JES Google Drive is not authorized. Staff are directed not to allow AI tools to save files to organizational drives. | Prevents unauthorized data exposure across shared drives. Pending formal governance decision ensures intentional rather than incidental integration. |
| AI in hiring | AI tools are permitted for drafting job postings, developing interview questions, and producing onboarding materials. Screening and shortlisting of candidates by AI is prohibited. | Consistent with JES's equity commitments and applicable employment standards across BC and international program jurisdictions. |
| Meeting transcription and records | AI-generated meeting records must be reviewed by the meeting organizer before circulation. All participants are informed before transcription begins. | Ensures accuracy of organizational records and protects information discussed in internal meetings. |
| Paid versus free AI tools | JES uses paid tiers of AI platforms (Claude Team, Google AI for Nonprofits). Paid accounts do not use organizational data to train AI models. Use of free AI platforms with JES data is not permitted. | Provides contractual assurance that JES program data does not contribute to AI model training. Reduces risk of inadvertent organizational data disclosure. |
Privacy Law and Standards
JES's primary privacy obligations in BC are governed by the Personal Information Protection Act (PIPA), which applies to BC nonprofits collecting personal information. JES's employee manual includes a dedicated Employee Privacy Policy (B5) that governs how personal information about staff is collected and handled.
PIPEDA (the federal Personal Information Protection and Electronic Documents Act) does not typically apply to nonprofit organizations that do not engage in commercial activities. JES voluntarily aligns to PIPEDA's 10 fair information principles as a best practice standard for responsible data governance, regardless of whether the Act applies.
For internationally funded programs, JES’s data handling obligations are additionally shaped by the terms of contribution agreements with federal government funders and the data protection expectations of host country partners. Our AI governance framework is designed to be consistent with all of these obligations simultaneously.
Ask AI: Design Principles
Ask AI is JES's purpose-built legal information tool. Its design reflects a deliberate governance approach:
- Ask AI provides legal information, not legal advice. It does not make moves in adversarial legal situations on behalf of users.
- All content in Ask AI is verified by qualified JES staff. The tool does not draw on general AI-generated legal content or produce unverified outputs.
- Ask AI is designed as a preparation tool: it helps people understand their rights before engaging with a legal professional, navigator, or advocate.
- No personally identifiable information about users is required to access Ask AI. The tool is designed to serve people who are not yet ready to speak to a person about their situation or want information outside of our call hours.
The tool opens a door. Navigators make sure it is the right door. The expert walks through it with them.
Ongoing Governance and Learning
JES treats AI governance as a living practice, not a one-time policy exercise. Our approach to ongoing governance includes:
- AI governance is a standing agenda item at Executive Leadership Team meetings, ensuring regular review of emerging issues, tool changes, and staff experience.
- We are establishing a shared organizational learning space so staff across all offices and programs can share AI experiences, flag concerns, and stay current with rapidly changing tools.
- Our AI policy documents are reviewed regularly and updated as tools evolve and new governance questions emerge. Staff are notified of material changes.
- Middle management guidance has been developed to help program directors support staff using AI, address quality concerns, and ensure that efficiency gains from AI are channelled into expanded program impact rather than reduced staffing.
JES acknowledges that AI governance is a field in rapid development.
Questions and Further Information
Funders and partners with questions about JES's AI governance approach are welcome to contact:
Carmen Zabarauckas, Ph.D.
Chief Executive Officer, Justice Education Society
ceo@justiceeducation.ca
A note on this document
This summary reflects JES's AI governance framework as of June 2026. It is intended as a high-level overview for funders and partners. The full internal policy (B14: Artificial Intelligence and Data Protection) is available on request. This document will be updated as our governance framework develops.