Artificial intelligence (AI) has long been reshaping the way we live and this summer marked an important stride in Canada’s path to an even more digitally-driven economy and government.
From Jan. 19 to Sept. 21, Minister Navdeep Bains of Innovation, Science, and Economic Development (ISED) led a series of roundtables as part of the National Digital and Data Consultations – an initiative designed to determine how the country can move forward as a leader in digital innovation, while also maintaining the integrity of privacy rights and consent.
Though Minister Bains has expressed his enthusiasm with how technological advancements – especially AI – can position the country as a global leader in innovation, critics like national Privacy Commissioner Daniel Therrien have voiced their concerns about the road we seem to be heading to. “At a time when new and intrusive targeting techniques are already influencing democratic processes and data analytics, automated decision-making technologies and artificial intelligence are raising important ethical questions that have yet to be answered, Canadians need stronger privacy laws, not more permissive ones,” Therrien told Teresa Wright of The Canadian Press.
Therrien’s stance is reflective of the current climate: as developments in AI continue to shift the technological landscape, issues surrounding privacy and consent have also grown. To address concerns with the responsibility that comes with AI, Minister Bains, along with Scott Brison, President of the Treasury Board and Minister of Digital Government, noted on TechVibes that the summer-long consultations included a diverse group of attendees from academia, civil society, and industry. Moreover, in advance of the consultations, the ISED stated that the diversity mandate means that groups representing visible minorities will also be consulted with.
Data retrieved through an Access to Information Request by the BC Freedom of Information and Privacy Association revealed that over 65% of the people consulted came from the industry sector, academia stood at 19%, while civil society was represented by 9% of the attendees. Another 5% of the reported attendees represented local government departments of their respective cities.
Held in 16 cities across Canada, the conferences included a session in Silicon Valley in the United States, as well as phone calls with Indigenous leaders from the First Nations Technology Council and from the First Nations Information Governance Centre. Beyond the calls, only Winnipeg, Whitehorse, and, Toronto included groups dedicated to First Nations causes. Additionally, only Winnipeg and Calgary had groups that represented the inclusion of women in the tech industry, while only Vancouver had groups representing immigration. Attendees for the Silicon Valley roundtable predominantly came from the industry sector with only one representative from academia. The consultation south of the border surprisingly featured no civil society groups, despite the proximity of groups such as the Electronic Frontier Foundation (EFF) and the American Civil Liberties Union (ACLU) of Northern California.
Diversity may not always be the first issue to be addressed regarding AI responsibility, but Minister Bains and Minister Brison both reiterated on TechVibes that bias in AI systems is a matter that their respective departments are looking very closely at. They indicated that the issue will not be seen as solely a cause for concern, but also as a “potential for learning,” writing that since AI algorithms are comprised of real-life data, they can then be repurposed to flag human biases.
It is the complex logic that can lead to convoluted, bureaucratic messages like this:
“You may have heard media reports about bias found in some AI systems. This is a subject we are closely studying, both as an area of concern and of potential learning. We’re exploring automated methods to scan for these biases, and flag decisions for human reviewers if they are detected. But, by the same token, AI can help us to identify and avoid bias. Because AI systems are fed with real-life data, AI can actually help reveal existing human biases so we can account for them.”
Essentially, in order to maximize the capabilities of data-driven systems that deal with machine-learning and decision-making, more data is needed to be collected.
I am no Luddite. I can be fascinated and engaged with the latest tech development just as much as anyone. I like that I can have an ongoing crossword puzzle app on my smartphone, or that I can access my fantasy basketball team from just about any internet-connected device – I am very much a fan of technology as a tool. But what we are seeing here is much bigger, I dare say even dangerous. We are already well aware that data regarding online habits and behavior have been manipulated for political means, and Minister Bains said it himself, “to say that AI is coming would be to miss the fact that it’s already here.” What we don’t have are the laws and protections that will render our privacy rights impervious to the seemingly limitless capacity of AI.
Therrien’s words to the ETHI committee ring an ominous, yet rightfully cautious bell. “Individual privacy is not a right we simply trade off for innovation, efficiency or commercial gain. No one has freely consented to having their personal information weaponized against them…. Similarly, we cannot allow Canadian democracy to be disrupted, nor can we permit our institutions to be undermined in a race to digitize everything and everyone simply because technology makes this possible.”
Canada has a bridge to cross when it comes to Artificial Intelligence. Some of us are already on the other side, and that should not mean that we should allow follow suit. Last I checked, that was what democracy was all about.
BC FIPA has filed more ATI requests on this matter and we will report on our findings as the process moves.
Carlo Javier is the Community Awareness and Outreach Coordinator at BC FIPA. He holds a Bachelor’s Degree in Communication Studies from Capilano University.