At Statistics Canada, we are always at the forefront, implementing leading-edge tools, technology and methods in what we do. And the collection, handling, processing and tabulation of massive amounts of data from numerous sources has always depended on the responsible use of those latest technologies and developments in software.
The agency continuously incorporates the use of the latest advances in ways that protect the privacy, confidentiality and quality of the outputs that bring greater value and efficiency to our operations and processes. Because being responsive to what Canadians need, also means being responsive to how we use these new technologies.
As artificial intelligence (AI) and machine learning (ML) continue to evolve and provide organizations around the world exciting opportunities to improve services, Statistics Canada has also begun exploring the potential benefits. But ensuring we use this technology responsibly and ethically is of paramount importance.
Adopting new technologies like AI and ML offers unique advantages when used with robust frameworks and good governance and does not come at the expense of how we safeguard your data. We are attentive to any risks associated with new processes and our experts continue to evaluate the benefits offered by any new technology with the necessary controls to ensure responsible application.
Whether we collect data by using AI, administrative data, or surveys, one thing remains certain—Statistics Canada continues to operate in accordance with governing instruments and frameworks, while ensuring your information remains private, secure and confidential. Visit our Protecting your privacy page to learn more, and check out Collecting your data to learn about the many ways Statistics Canada collects data and how it benefits you.
Visit our website to learn more about Statistics Canada's framework for responsible machine learning. To read more on the Government of Canada's guidelines on the responsible use of AI, please see the Guide on the use of Generative AI.
Artificial intelligence for prediction and data collection
There are many ways in which AI facilitates data collection and makes predictions. Modernization efforts underway at Statistics Canada will reduce response burden on farmers responding to agriculture surveys with innovative data-collection methods such as image processing.
For example, the Agriculture Statistics Program is using supervised ML for the following projects:
- In 2019, the agency implemented a model-based method for Manitoba to produce crop yield predictions using longitudinal satellite observations of vegetation levels as well as regional meteorological measures. See: Use of Machine Learning for Crop Yield Prediction.
- During the 2021 Census of Agriculture, Statistics Canada carried out a ML project to determine the total surface area of greenhouses in Canada using satellite imagery. See: Greenhouse Detection with Remote Sensing and Machine Learning.
- An agency project used existing satellite imagery from 2018, 2019 and 2020 to determine which crops were growing in agricultural land. The results were 95% conclusive. See: Extracting Temporal Trends from Satellite Images.
Artificial intelligence for automation and data analysis
Statistics Canada is using AI techniques in the following ways to analyze, categorize and group data:
- 2026 Census Chatbot: Statistics Canada is currently developing a new chatbot for the 2026 Census to improve communications and support services provided to Canadians. This tool is being developed using a combination of open-source and SaaS (software as a service) ML solutions. The service will provide Canadians with timely, accurate, and automatic responses to frequently asked questions and will provide a pathway to a live agent, if needed. This chatbot will only use responses that have been written and reviewed by subject-matter experts from Statistics Canada to ensure accuracy.
- Canadian Coroner and Medical Examiner Database: Statistics Canada works with provinces and territories to collect data for the Canadian Coroner and Medical Examiner Database. Since each province and territory has its own method for classifying data, ML is used to organize the collected data into coherent datasets. Analysts then assess the datasets for patterns of death over time. Detecting trends in mortality allows medical examiners and coroners to understand growing hazards. These are reviewed and validated by experts within the agency to ensure the mitigation of biases and preserve the quality of our outputs before being flagged as public health threats to Canadians.
Governance frameworks for using artificial intelligence
Statistics Canada is committed to using AI in a responsible and ethical manner. Our processes are constantly evolving to adapt to the opportunities and challenges associated with AI and other technologies. We respect our commitment to Canadians, and we operate in accordance with governing instruments and frameworks that guide its responsible use, including:
- Protection of confidentiality in accordance with the Statistics Act: The Statistics Act ensures that the information provided to us is kept confidential. Maintaining the trust of Canadians and protecting their personal information is always a priority for the agency. See: Privacy and confidentiality for more information.
- Cybersecurity assessments: Every new project that uses AI undergoes a cybersecurity risk assessment of all the applications, systems and software that will be used to process sensitive and personal information. This assessment identifies and manages any potential threats, vulnerabilities and risks associated with processes, technologies (software vulnerabilities), people and documents.
- Governance instruments: We comply with all Government of Canada directives on the Responsible use of artificial intelligence, including the Directive on Automated Decision-Making.
- Governance committees: Statistics Canada works with many governance mechanisms that approve data-collection projects. The Advisory Council on Ethics and Modernization of Microdata Access complements the guiding role played by the Canadian Statistics Advisory Council. For more information on ethical reviews, watch the Data ethics: An introduction and Data ethics part 2: Ethical reviews videos.
- Principles of necessity and proportionality: Privacy and confidentiality are central principles in every step of a project involving data collection. The Necessity and Proportionality Framework was developed in partnership with the Office of the Privacy Commissioner of Canada, in full accordance with the Statistics Act and the Privacy Act, to balance society's need for official statistics, with the need to reduce the response burden on Canadians, all while protecting their privacy.;
- Foundations of ethical considerations: Six guiding principles: These principles ensure privacy protection optimization and production of information when designing a data-gathering approach. They include benefits for Canadians, privacy and security, transparency and accountability, trust and sustainability, quality, as well as fairness and do no harm.
- Framework for responsible machine learning: To guide the ethical use of data processed using ML techniques, Statistics Canada has developed the Responsible use of machine learning at Statistics Canada framework for statistical programs and projects that use ML algorithms.
Visit Data science projects to learn more about the use of AI at Statistics Canada in areas such as natural language processing and image classification. Statistics Canada continues to explore the use of AI and ML as a solution to facilitate data collection, to categorize and make safe and effective predictions about data, and to enhance the value of the projects within the agency.
Learn more about the basic concepts about ML and AI with Machine learning: An introduction.