AI, Virtual Care, and Accreditation

Strengthening Indigenous Health Systems Through Responsible Innovation

Sidney Shapiro, PhD
Assistant Professor of Business Analytics
Dhillon School of Business
University of Lethbridge

Digital health and Indigenous cultural presence

Context: Why This Conversation Now?

Severe health human resource shortages across Canada
Colonial legacies, underfunding, and barriers to care have contributed to health disparities.
Higher prevalence of chronic conditions in many First Nations communities
Geographic and infrastructure barriers to care access
Increasing digital infrastructure in rural and remote regions
Accreditation focus on quality, safety, and continuous improvement

Core Question: How can AI-enabled virtual care strengthen quality and accreditation readiness?

A Brief History of AI

Early Foundations (1950s–1980s)

1950: Alan Turing proposes the "Turing Test"
1956: Dartmouth Conference coins "Artificial Intelligence"
1960s–70s: Expert systems and rule-based AI
1980s: First AI winter—limited computing power

Modern Era (1990s–Present)

1990s: Machine learning gains traction
2000s: Big data enables better training
2010s: Deep learning breakthroughs (ImageNet, AlphaGo)
2020s: Large language models transform productivity

Key milestone: The convergence of massive datasets, powerful GPUs, and improved algorithms in the 2010s enabled today’s AI revolution—including tools that support health and virtual care.

What Are Large Language Models? (Generative AI)

What they are

LLMs are AI systems trained on vast amounts of text to understand and generate human-like language:

Neural networks with billions of parameters
Trained on internet-scale text data
Predict the next word based on context
Can summarize, draft, and support documentation

Growth and adoption

2023: ChatGPT reaches 100M users in 2 months
2024: Widespread use in enterprise and health
2025: AI tools become standard in workflows

                In health: Gen AI can assist with clinical notes, triage support, and reporting—but it works best when grounded in your own data and used with human oversight.
            

Tim Hortons

AI in Health: What We Actually Mean

AI in health typically includes:

Clinical decision support tools
Risk prediction models
Automated triage systems
Remote patient monitoring
Documentation automation
Population health analytics

This is not science fiction. These are operational tools already in use.

Data, ETL, and Finding New Patterns

Many of the gains from AI in health depend on how data is gathered and prepared. ETL—Extract, Transform, Load—describes the pipeline: pulling data from source systems (E), cleaning and structuring it (T), and loading it into a place where it can be analyzed or used by AI (L).

In your context: Data might come from visits, program reports, community surveys, or accreditation evidence. When it’s extracted and transformed consistently, it becomes usable.
AI on that data: Once data is in a structured form, AI can look for patterns—trends in service use, gaps by community or program, early signals of risk—that are hard to see in spreadsheets or paper.

Bottom line: Gathering and preparing data well is what makes “AI finding new patterns” possible. Doing it in ways that respect data sovereignty is what makes it responsible.

The Health Human Resources Crisis

Challenges:

Provider shortages
Burnout
Recruitment and retention barriers
Increased chronic disease burden
Long wait times and travel requirements

Key point: AI-enabled virtual care does not replace providers. It reallocates time and increases system capacity.

Where AI Can Bridge Gaps

1. Virtual Triage and Access

AI-assisted intake systems
Symptom prioritization
Directing patients to appropriate services

2. Chronic Disease Management

Remote monitoring for diabetes, hypertension
Risk alerts for early intervention

3. Documentation Burden Reduction

Automated note generation
Structured data extraction
More provider time with patients

Virtual health consultation in a rural community

AI in Use Today: Real Examples

AI is already supporting clinicians and patients in concrete ways:

Example: Clinical note-taking (e.g. Heidi)

AI tools can listen to the conversation between provider and patient and generate structured clinical notes in real time—freeing the provider to focus on the person, not the keyboard.

Reduces documentation burden during and after visits
Improves accuracy and consistency of records
Supports accreditation-ready documentation

Tools like these are in use today; the question is how to adopt them in ways that align with your governance and standards.

Accreditation Lens: Where AI Actually Touches Your Work

The pressure point is often evidence at review time—showing that improvement is systematic, that documentation is consistent, and that risk is identified before it becomes a finding.

Documentation that reviewers can trust: AI-assisted notes and summaries only help accreditation if they’re accurate, verifiable, and aligned with what actually happened. Otherwise they create risk.
From “we do it” to “we can show it”: Dashboards and pattern-finding can surface the right data at the right time—so you’re not reconstructing the story after the fact.
Consistency across sites and programs: Standardized extraction and reporting can reduce the gap between locations that document well and those that struggle, so the organization presents a coherent picture.

Bottom line: AI is relevant to accreditation not because it’s new technology, but because it can either close the gap between doing the work and demonstrating it—or widen that gap if adoption is poorly governed. Your role is to decide which.

You are not Google

Most organizations are not looking for build technology for technology's sake. You are looking to improve your workflow and your organization to meet the needs of your community.

Using AI for Logic Models, Theory of Change, Gantt Charts, M&E Plans, and Related Evaluation Tools

AI can be used to help with the creation of logic models, theory of change, Gantt charts, and related evaluation tools.

AI as a Quality Improvement Tool

AI strengthens:

Evidence-informed decision making
Outcome measurement
Service equity monitoring
Pattern detection across programs

It supports the move from reactive to proactive care.

Risk and Governance Considerations

Accreditation requires safeguards.

Key issues:

Data sovereignty and governance
Algorithmic bias
Transparency and explainability
Informed consent
Procurement and vendor accountability
Vendor lock-in and data ecosystems

Indigenous Innovation in AI

Indigenous Pathfinders in AI – Mila (Montreal)

Training First Nations, Inuit, and Métis AI leaders
Building internal capacity rather than external dependence

First Nations Technology Council

“The Missing Code” initiative
Aligning AI adoption with Indigenous rights and governance

First Languages AI Reality (FLAIR) – Mila

AI for Indigenous language revitalization
Connects technical capacity with community-led language work

First Nations Information Governance Centre (FNIGC)

Foundation for how health and AI systems should be governed in First Nations contexts

Indigenous leaders reviewing digital governance

Indigenous Pathfinders in AI (Mila)

Montreal-based program building Indigenous leadership in AI:

Trains First Nations, Inuit, and Métis individuals to lead in AI development—not only to use it
Creates internal capacity so communities are not dependent on external vendors
Ensures Indigenous perspectives shape how AI is built and governed

First Nations Technology Council: The Missing Code

British Columbia initiative ensuring AI adoption respects Indigenous rights:

“The Missing Code” addresses the gap between mainstream AI development and Indigenous rights and self-determination
Works so that AI adoption in BC aligns with First Nations governance and data sovereignty
Bridges technology policy and Indigenous leadership

Language & Cultural Preservation

AI and language technology are being used to revitalize Indigenous languages and preserve traditional knowledge:

Speech recognition and language models for Indigenous languages
Digital archives of traditional knowledge
Canada: The NRC Canadian Indigenous Languages Technology Project works with Indigenous communities to develop speech- and text-based tools (e.g. ReadAlong Studio, predictive text, verb conjugators) for over 25 languages; software is released to communities as open-source.

Language technology and intergenerational collaboration

AI for Holistic & Culturally Informed Care

AI can support care that reflects community values, connection to land, and whole-person wellbeing—when it is designed and trained with those goals in mind.

Whole-person care: Models can be trained to recognize patterns that matter in Indigenous health—family and community context, connection to culture, social determinants, and spiritual wellbeing—not only clinical symptoms.
Community-specific patterns: AI trained on data that reflects First Nations populations can learn unique patterns of health, resilience, and need that generic models miss.

Key idea: AI is not culturally neutral. When trained using data and priorities that reflect First Nations peoples’ unique patterns and values, it can support more holistic, culturally safe, and effective care—and better alignment with accreditation and community expectations. Systems can be designed to include cultural data during training or using fine-tuning techniques to add to an existing model.

Practical Applications: Knowledge & Environment

AI is used to blend traditional knowledge with scientific data:

Qikiqtaaluk Corporation (Arctic): Environmental management and decision support using both local knowledge and data
Mapping new fishing locations by combining traditional ecological knowledge with scientific and sensor data
Supporting land and resource management in ways that respect Indigenous governance

These applications show how AI can integrate community knowledge systems—a principle that applies to health and accreditation as well.

What This Means for Health Accreditation

When aligned with your existing quality and governance expectations, AI can help close the gap between day-to-day practice and what reviewers need to see.

AI adoption can support:

Documentation quality
Standardized care pathways
Improved reporting
Continuous quality monitoring
Staff workload sustainability
Create apps that support the work of the organization

The payoff is not only efficiency—it is clearer evidence of systematic improvement and risk management for your organization.

It can strengthen compliance readiness if governed properly.

Strategic Questions for Accreditation Leaders

How could AI reduce risk in your service delivery?
Where are documentation burdens slowing quality improvement?
What data do you wish you had in real time?
How can Indigenous data governance be embedded from the start?
What standards would you require before approving AI tools?
Where does AI make sense in your organization?

Practical Starting Points

Think about the work you do. What can AI help you do better? What are the workflows that could be improved with AI?

Low-risk entry points:

Documentation automation pilots
Predictive dashboards for chronic disease
AI-assisted quality monitoring
Virtual triage support

Start small. Measure impact. Align with accreditation metrics.

Closing Thought

AI is not a solution on its own.

It is a systems amplifier.

When aligned with Indigenous governance, accreditation standards, and quality improvement, it can:

Expand capacity
Improve safety
Reduce burden
Support equitable access

Discussion

Open conversation:

What feels promising?
What feels risky?
Where do you see immediate opportunity?
What would accreditation need to see to support AI integration?