On June 7th, 2018, Professor Elizabeth Dubois, member of the Centre for Law, Technology and Society, appeared before the House of Commons’ Standing Committee on Procedure and House Affairs (PROC).
The Committee is currently conducting a review of Bill C-76, an Act to amend the Canada Elections Act and other Acts and to make certain consequential amendments.
Professor Dubois is a Member of the Centre, and Assistant professor at the Department of Communication, Faculty of Arts of the University of Ottawa.
- See below for Prof. Dubois speaking notes.
- The audio recording of the hearing is available here.
Hello and thank you for inviting me to speak today regarding Bill C-76. My name is Elizabeth Dubois and I am an Assistant Professor in the Department of Communication at the University of Ottawa. My research focuses on how people access and share political information and the role of digital media such as social media platforms and search engines in that process. For example, with Dr. Fenwick McKelvey at Concordia University I co-authored the first, and I believe only, publicly available and academically rigorous report on the state of political bots in Canada as part of the University of Oxford’s computational propaganda project.
Today I want to draw your attention to three key aspects of C-76 in my opening remarks.
1. Computational approaches to voter suppression
2. Technology and platform companies
3. Political party privacy policies
First, based on evidence from recent elections and referenda internationally, we know that individuals and groups are experimenting with computationally supported political tactics which could lead to voter suppression.
These techniques could include creating automated social media accounts (non-human accounts called bots), creating fake accounts and troll accounts (run by humans but which are not necessarily representative of actual voters), and targeted advertisement strategies which could involve quickly removing ads so they are hard to trace.
By using computational approaches and automation it is possible to amplify and spread information very quickly. It is also possible to dampen messages and supress ideas. This can be used for obvious and explicit forms of voter suppression such as telling people to go to the wrong polling place – one could imagine a bot driven version of the Robocall scandal, or for more covert forms of suppression such as creating an environment of distrust in the electoral system or encouraging political apathy – which could be done via a chatbot which is a basic form of artificial intelligence. Notably, most research currently considers the role of political bots on social media but increasingly tools such as WhatsAppp and other instant messaging applications are being employed. Voter suppression in these contexts is even harder to identify.
This is clearly against the spirit of the law but not explicitly addressed. Nor are there adequate mechanisms in place to prevent or identify these practices. A requirement to register use of automated techniques (which includes emerging forms of artificial intelligence) for communicating with the electorate would be a valuable addition to this legislation.
Note, I say register and not ban, I believe there are valuable and legitimate uses of automated techniques for communicating with the electorate that should not necessarily be discouraged.
Considering my second point, the role of platforms such as major social media companies and search engines could be better addressed in C-76. The bill requires organizations to not knowingly sell election advertisements to foreign entities which will affect platform companies. However, the bill ignores the substantial role platforms play when it comes to enforcing many aspect of Canadian election laws.
For example, the low cost of online advertisement and ability to micro-target means that hundreds of versions of advertisements can be made and distributed. They are hard to track and therefore it can be difficult to establish if and when illegal activities are happening such as voter suppression or advertisement spend which exceeds spending limits, is purchased by foreign entities, or is purchased by unregistered third-parties.
Platform companies, having been confronted with this problem elsewhere such as the US, have started to create advertisement transparency tools. However, this is voluntary and could be changed at any moment if not required legally. This poses a significant risk to Canadian elections because platforms make decisions in an international and commercial context which does not necessarily align with the needs of Canada’s democracy.
Finally, C-76 requires political parties to make a privacy statement about protecting personal information of individuals. The proposed legislation does not include any form of audit or verification that the policy is adequate, ethical, or being followed. There are no penalties for non-compliance. There are no provisions that permit Canadians to request their data be corrected or deleted, as is the case in other countries.
It is fair to say this issue is much broader than elections but the fact that political parties are not covered by PIPEDA and the fact that elections are fundamental to the functioning of our democracy means it is an issue that cannot be ignored.
Ultimately, I think there are useful aspects to this bill but there are substantial concerns regarding computational approaches to voter suppression, the role of technology companies and platforms, and privacy which I hope will be considered in more detail.
Thank you for your time and I look forward to questions.