AI-Assisted Document Summarization for Veterans Disability Claims
Healthcare · Eligibility & Public Benefits
What it collects that can identify you
- Veterans' medical assessments, disability records, and personal client files stored in the VAC Client Services Delivery Network and GCCase. This includes health information that constitutes sensitive personal data.
- Client records from the VAC Client Services Delivery Network and GCCase, which identify specific veteran clients and are linked to their accounts and benefit histories.
- Run by
- Veterans Affairs Canada (VAC)
- Where
- No fixed location
- Kept
- Not stated by the Helpful Places.
- Shared with
- Accountable organization
- Your copy
- You cannot see the data it holds about you. What you can do
What it is for
This system uses generative AI to read and summarize personal records and medical assessments held by Veterans Affairs Canada, producing evidence bundles that help adjudicators decide disability claims. It is currently being tested and is not yet in full production. Veterans' personal and medical information is processed, and staff are informed that AI is in use.
What it collects and what happens to it
Data taken in
- Veterans' medical assessments, disability records, and personal client files stored in the VAC Client Services Delivery Network and GCCase. This includes health information that constitutes sensitive personal data.
- Client records from the VAC Client Services Delivery Network and GCCase, which identify specific veteran clients and are linked to their accounts and benefit histories.
Processing
- Azure OpenAI large language models are used to synthesize extracted document content into coherent written summaries and evidence bundles for adjudicators.
What it does
- Uses generative AI (Azure OpenAI) to produce written summaries of client documents and evidence bundles. The generated content is advisory — a human adjudicator reviews and decides on each disability claim.
- Uses Document Intelligence capabilities including handwriting recognition, general document analysis, and custom extraction and classification models to extract structured information from scanned forms and handwritten records.
Outputs
- A concise evidence bundle and written summary of the veteran client's documents and medical assessments, produced for use by a disability adjudicator. The output is a synthesized document, not a copy of the source records.
- The evidence bundle serves as an advisory input to the adjudicator's determination — it presents and organizes information to assist the human decision-maker but does not itself constitute a binding eligibility decision.
Run by
- Federal department responsible for delivering benefits and services to Canadian veterans. Operates the Claims Summary Tool to support disability adjudication.
Built by
- Technology vendor providing Document Intelligence and Azure OpenAI capabilities used by the Claims Summary Tool. The system was developed in part by the VAC AI Development team using Microsoft services.
Kept for
Not stated by the Helpful Places.
Shared with
- Outputs are available to Veterans Affairs Canada adjudicators (GC employees) who use the evidence bundles to support disability claim decisions. Access is limited to authorized departmental staff.
- Veteran clients do not have direct access to the AI-generated summaries produced about their own records. The output is an internal adjudicator-facing tool; veteran access to underlying source records is governed by separate VAC access procedures.
Stored
Not stated by the Helpful Places.
How to read the colours
Can it identify you?
- Anonymized data
- Data about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so this system can’t reasonably tie a record back to an individual.
- Pseudonymous data
- Each person’s data is tied to a token (hash, ID, template) that lets this system recognise the same person across events, but the token itself doesn’t reveal a name. Reidentification is possible with extra information.
- Identifiable data
- The data either contains a direct identifier (name, address, account name, recognisable face or voice, plate number) or carries a token this system uses to look up legal identity during processing.
Who completes the loop?
- Human decides
- This mode suggests; a person decides what to do next. The AI is always advisory — a human is in the loop on every decision. Example: a triage tool ranks cases for a clinician who chooses which to see first.
- Human executes
- This mode decides; a person carries out the result. Example: an optimizer plans the day’s trash-collection routes, and drivers run them.
- Autonomous
- This mode decides and acts on its own. No person reviews each decision or carries out the resulting action.
Definitions from the DTPR standard. Amber is about your data, violet about who decides. The fuller the shape and the deeper the colour, the more identifying the data or the less a person is involved.
- AI registerGovernment of Canada Algorithmic Impact Assessment Register — Claims Summary Tool (CST)Veterans Affairs Canada, AI Register entry 2526-VAC-ACC-002.
- AI registerGovernment of Canada AI Register — CST entry
- AI registerGovernment of Canada AI Register — CST entry
- Register entryPublished by the Helpful Places. Reference 3ca0d518. This disclosure was drafted with AI assistance.Schema: ai@2026-05-06-beta
What you can do
Ask about this system
Questions go to the Helpful Places, not the vendor.
Your rights
- Right to Algorithmic TransparencyVeterans Affairs Canada has disclosed that AI is in use within the adjudication support process. The system is listed on the Government of Canada's public AI register, providing transparency about its general purpose, capabilities, and data sources.
- Right to a Human ReviewAll disability adjudication decisions are made by human adjudicators at Veterans Affairs Canada. The AI system produces advisory summaries only; no benefit eligibility decision is made solely by the AI. Veterans retain the right to have their claims reviewed and decided by a qualified human adjudicator.
Risks and safeguards
- Reputational harmInaccurate or incomplete AI summaries could misrepresent a veteran's medical or service history, leading to unfair adjudication outcomes and reputational or financial harm to the veteran.Safeguard: The system is currently in development and being tested with adjudicators to verify that evidence bundles are comprehensive and accurate before broader deployment. Human adjudicators review all AI-generated summaries before any claim decision is made.
- Civil liberties harmProcessing sensitive medical and personal records of veterans through a third-party AI platform (Microsoft Azure OpenAI) raises concerns about data sovereignty and the right to privacy under Canadian law, particularly regarding cloud data handling.Safeguard: Data sources are restricted to VAC's internal systems (Client Services Delivery Network, GCCase); the register confirms personal information is involved and AI use is disclosed to system users (GC employees).