Five ways the DHDP is promoting the ethical use of AI to improve health outcomes for Canadians

Every time a person interacts with the healthcare system millions of data points are generated. Similarly, the more we study different diseases, the more data we generate on them.

These vast amounts of data hold answers to important questions that may help us improve care for millions of people. But humans may never be able to analyze this growing data on their own.  

That’s where AI comes in.

AI can quickly analyze data to reveal earlier signs of disease, support more personalized treatments, and accelerate discoveries that could improve countless lives. Precision medicine, where treatments are tailored to the individual, relies on this type of analysis. In fact, the more data AI has, the better its analysis, meaning that for this to create the most value, data sharing and AI analytics deployment across institutions and across Canada is imperative.

But the technology itself is only part of the story.

The real story is about building a healthcare ecosystem in which Canadians can feel confident that innovation and privacy go hand in hand. And that means defining the ethical use of AI in digital health data research...What is ethical use of AI?

AI scare stories abound these days, and rightfully so. From data breaches and agents gone rogue through stereotype and bias being blatantly promoted, to the environmental impact of data centres. All these are valid concerns about a new technology with the capacity to have a huge impact on the world.

But is there actually a safe and ethical way to use AI, especially with digital health data?

Ethical AI is sometimes described as a set of principles, guidelines, or policies. There are actual definitions for this globally, with practical implementation guided by frameworks like the EU Ethics Guidelines for Trustworthy AI or national policies such as the Canadian Responsible Use of AI, for example. These broadly define ethical use under the core principles of ensuring that systems are fair, transparent, and designed to benefit people while minimizing harm.

At its core, AI must respect human rights and human dignity.

In practice, however, ethical use of AI is something much more tangible. It is the decision to design systems that put people first. It is building safeguards before problems arise. It is recognizing that public trust is not an obstacle to innovation but a prerequisite for it.

Dilemma: innovation or privacy?

So how can we make sure AI is a partner in improving health for all Canadians? We’re facing a real ethical dilemma: innovation vs. privacy?

In order to realise Canadian digital health data value in precision medicine research, we need to share the data and analyze it effectively, which means deploying AI in various forms. The two key factors to consider are how we share health data and then how we deploy AI tools. There are very real concerns about preserving patient privacy, confirming consent, and establishing good governance on making data available to AI analytics.

One way forward would be to severely restrict data sharing, keeping it in-house and open only to researchers at the institute where the data are generated. But with this, data cannot be pooled and analyzed as a larger cohort, and this is especially harmful to rare disease and cancer research where cases are isolated and not part of a bigger picture. The ability to look into larger data cohorts brings better results; bigger is definitely better for AI-based analytics.

Ethical use of AI in action; could it exist?

But pooling data means moving it? Not necessarily.

In a federated learning model, data sharing exists in a discovery format, with only the analytical computation moving, not the data. Raw data stays put but it’s findable by researchers across Canada; the federated learning infrastructure not only activates AI analytics and data discovery, but it also retains patient data under secure institutional privacy and ethics governance.

The Digital Health & Discovery Platform (DHDP) is being built along those lines with pan-Canadian collaboration for usability and also to ensure data privacy for ethical use of AI.

Ethical AI in healthcare is not just about algorithms; it's about building privacy-preserving infrastructure, ensuring human oversight, improving equity, and using data responsibly to deliver better outcomes for patients across Canada.

5 ways DHDP is promoting the ethical use of AI

Below are five projects supported by the Digital Health Innovation Fund (DHIF) that are promoting the ethical use of AI to improve health outcomes for Canadians.

Using AI-powered imaging to detect bone cancer early and improve patient outcomes in real time

Project: An AI-Driven Multimodal Spectroscopy Platform for Real-Time Bone Cancer Detection

Today, as when Terry Fox faced his osteosarcoma diagnosis, surgeons often have limited ways to know if all cancer has been removed, especially in hard bone tissue. If cells from these aggressive tumours are left behind after surgery, the disease can return. Repeated surgeries may be required; limb amputation is probable. Metastatic spread to other areas of the body can be devastating.

The project team will develop a handheld device and AI-driven spectroscopy platform for real-time bone cancer detection. Using AI and machine learning, this device will learn to detect cancer from among the complex tissue signals encountered during an operation and then give instant feedback to the surgeon.

Using privacy preserving synthetic data to connect hospitals across Canada and unlock faster insights to improve patient care

Project Name: Building a Synthetic Data Network in Canada

Initially focusing on lung cancer, the project will standardize hospital data and then generate a synthetic dataset. This data carries all the statistical characteristics of the original but protects personal health information. Since synthetic data retain only clinical information and not personal details, they can be used to train AI tools and help analytical models learn, giving valuable insight on treatment pathways, diagnostic work ups, and patient outcomes.

Project: EvidaHealth Foundation: National Registry Infrastructure Enabling FAIR, Federated, and Patient-Centered Real-World Evidence

Health data in Canada is often fragmented across provinces, hospitals, and systems. This makes it difficult to answer important questions about long-term outcomes, treatment effectiveness, and patient experience, particularly for complex conditions.

Building a secure Canadian skin-health dataset so AI can help doctors spot high-risk cases sooner.

Project: Use of Canadian Data, Including Indigenous Data, to Train AI Models on Skin Cancer and Other Dermatological Diseases

Precision medicine relies on analyzing big data sets for answers, and this often needs AI to pull out patterns that the human eye cannot. But the answers that AI pulls out are only as good as the data that it trains on. With missing data, results generated can be biased and missing impact for a whole section of the population.

A new project supported through the Digital Health & Discovery Platform’s Digital Health Innovation Fund is not only exploring how AI can help address these challenges but also making specialist health care for skin cancer accessible to remote and rural communities. 

Enabling faster, smarter vision care across Canada through real-world AI and secure data sharing

Project: Federated Learning for Eye Health: Advancing Vision AI Models Through the DHDP Data Infrastructure

There’s a growing challenge in health care; how to make full use of fragmented and unstructured data collected in everyday clinical settings. These data hold valuable insights that could support practitioners and patients for early detection of eye disease.

This collaboration between partners in British Columbia and Alberta gathers vision data from both rural and urban areas to develop AI-driven tools that can support clinicians in real time, reduce diagnostic delays, and improve access to care across diverse populations.

For all projects funded under the Digital Health Innovation Fund, please visit our Projects page.