Date and location:
June 27, 2024, 9:30 a.m. to 12:00 p.m.
Hybrid meeting
Microsoft Teams
The Ivan P. Fellegi Conference Room, 26th floor, R.H. Coats Building
100 Tunney's Pasture Parkway, Ottawa
Members:
- Chantal Bernier, Co-Chair, Global Privacy and Cybersecurity Group, Dentons Canada (Chair)
- André Loranger, Chief Statistician of Canada
- Len Garis, Researcher and Adjunct Professor, University of the Fraser Valley
- Ali Ghorbani, Cyber Security Chair, University of New Brunswick
Regrets:
- Robert Gordon, Executive Director, Canadian Cyber Threat Exchange
- David Robichaud, Ethics Professor, University of Ottawa
Statistics Canada guests/support:
- Josée Bégin
- Geoff Bowlby
- Pierre Desrochers
- Geneviève Jourdain
- Patrice Martineau
- Kathleen Mitchell
- Saeid Molladavoudi
- Nicolas Rahal
- Eric Rancourt
- Christos Sarakinos
- Jennifer Withington
1. Introductions and Opening Remarks
André Loranger, Chief Statistician of Canada
Chantal Bernier, Co-chair, Global Privacy and Cybersecurity Group, Dentons Canada LLP
Mr. Loranger began the meeting by welcoming members to the eighth meeting of the Advisory Council on Ethics and Modernization of Microdata Access, his first as Chief Statistician of Canada.
Mr. Loranger then presented his vision for Statistics Canada. He reiterated his intention to continue the modernization efforts started a few years ago by the organization, by making a slight change in emphasis. According to Mr. Loranger, it will be based on the following pillars: a clear strategic vision through Statistics Canada's Strategic Plan, quality infrastructure to support activities, competent and adequately trained teams and the positioning of the United Nations Fundamental Principles of Official Statistics as a central pillar guiding decisions.
According to Mr. Loranger, the organization will have to face significant challenges over the coming years, some of which will be discussed today. The Council will play an important role, providing sound advice and insight.
Ms. Bernier, for her part, congratulated Mr. Loranger on his new position and underlined the relevance of the subjects to be discussed during the meeting. She recalled the importance, for an organization such as Statistics Canada, of remaining relevant. To support her statement, she cited Canada's former chief statistician, Ivan Fellegi, who once told her: “Make sure your organization is always relevant. If you're excellent but not relevant, nobody will care.ˮ
2. Mis- Dis- and Mal-Information and Official Statistics
Pierre Desrochers, Director, Office of Privacy Management and Information Coordination, Statistics Canada
Mr. Desrochers spoke about the importance for an organization like Statistics Canada to curb misinformation by providing quality data, to maintain public trust and democratic integrity. As Mr. Desrochers indicated, this responsibility corresponds to Principle 4 of the United Nations Fundamental Principles of Official Statistics, Prevention of misuse.
Mr. Desrochers then explained the distinction between misinformation (false information, but the person who disseminates it believes it to be true), disinformation (false information and the person who disseminates it knows it) and malinformation (true information and the person who disseminates it does it with the aim of causing harm to a person, an organization, a country, etc.) Mr. Desrochers concluded by presenting some legislative measures put in place at the international level to prevent the abusive use of statistics, recalling however the limited scope of this type of intervention.
Council members were pleased with the presentation, emphasizing its great relevance in the current context. The Council strongly recommended that Statistics Canada play a more proactive role regarding the misuse of statistics and adopt a holistic approach, centered around three elements:
- Raising awareness and educating Canadians to help the population recognize the misuse of statistics and be able to respond adequately.
- Technology to detect the misuse of statistics.
- Intervention, to correct the identified issues. For this element, the Council suggested a proactive approach where areas of vulnerability would be identified quickly and quality statistics to inform debates would be produced and disseminated.
Finally, the Council recommended that Statistics Canada adopt an official strategy to combat the misuse of statistics, supported by the document Countering Disinformation: A Guidebook for Public Servants, produced by the Government of Canada, and to be as proactive as its mandate allows.
3. Navigating Responsible AI: Insights from Statistics Canada's Journey, with Focus on Generative AI Christos Sarakinos, Director, Data Science and Innovation Division, Statistics Canada
Saeid Molladavoudi, Senior Data Science Advisor, Data Science and Innovation Division, Statistics Canada
Mr. Molladavoudi presented Statistics Canada's work in generative AI and its commitment to responsible AI. As highlighted by Mr. Molladavoudi, the Government of Canada has put in place clear policies regarding AI and Statistics Canada has adopted a comprehensive framework including a governance structure, directives, seven core guiding principles, six pillars and many activities.
The members of the Council were pleased with the presentation and wanted to highlight the progress and the high-quality work carried out by Statistics Canada in this area. The use of AI offers many opportunities, including the production and publication of statistical data, but also presents certain risks.
Among the risks, Council members highlighted those related to data confidentiality, privacy, cybersecurity, false authoritative results and potential bias. For the latter, in particular, validation by a human being remains important. All these risks could greatly affect the organization's reputation and the public’s trust in it.
Council members also suggested adding a guiding principle related to the impact on individuals. The Council would like to have an update on the progress made at a future meeting, ideally during the next year.
4. How to ensure data confidentiality while offering data users fast and flexible access
Geneviève Jourdain, Director, Data Access Division, Statistics Canada
Ms. Jourdain presented Statistics Canada's current strategy regarding data access for external users, the challenges it represents as well as the proposed solutions. These include: (1) the possibility of obtaining access to microdata for a specific program rather than on a project-by-project basis and (2) the delegation of the data confidentiality vetting process to deemed employees of Statistics Canada.
Following questions from Council members, Ms. Jourdain clarified that these options would only be offered to federal partners, that Statistics Canada would retain the right to do the vetting itself when the situation warrants it, and that the data always remains in Statistics Canada's secure environment. Deemed employees would also receive thorough and appropriate training.
Still in response to questions from Council members, Ms. Jourdain added that these changes are motivated by an increase in the number of requests, by the need to modernize our practices and by the demand of our main partners.
The members of the Council were pleased with the presentation and the proposed solutions, which they believe adequately address the issues identified and maintain the security of people, data and infrastructure.
Council members recommended that Statistics Canada carry out more ex-post evaluations, to ensure that data are used as originally agreed.
5. Roundtable and Closing Remarks
André Loranger, Chief Statistician of Canada
Chantal Bernier, Co-chair, Global Privacy and Cybersecurity Group, Dentons Canada
Mr. Loranger and Ms. Bernier thanked the members of the Council as well as the participants. The Council agreed to include, at the next meeting, a session during which Council members will propose relevant issues that Statistics Canada could address in the short term.
Advisory council action items
Action Items
- Next meeting: fall 2024.
- Include a session during which Council members will propose relevant issues.
Recommendations for Statistics Canada
- Mis- Dis- and Mal-Information and Official Statistics
- Have an official strategy to combat the misuse of statistics and be as proactive as its mandate allows.
- Adopt a holistic approach, focused on three elements: (1) Education and awareness; (2) Use of technology to identify misuse; (3) Intervention, to correct the identified issues.
- Navigating Responsible AI
- Add a guiding principle related to the impact on individuals.
- Come back and present the progress made within a year.
- Data confidentiality and fast and flexible access
- Implement the new options proposed.
- Carry out more ex-post evaluations.