Expand domestic markets for home-grown applications
So far, our reputation has been acquired largely as a result of an early and sustained commitment by federal and provincial governments to fund pioneering research in AI and its theoretical foundations. This choice has paid huge dividends; it has enabled us to establish an impressive base of talent and research, and it has encouraged the presence of a deep culture and community of AI trailblazers and superstar scholars. Yet despite this investment in Canada’s AI R&D ecosystem, we have only begun to scratch the surface of our country’s potential. The next step is to ensure Canadian companies are benefiting from our innovations in AI.
We know that AI has the potential to improve productivity across sectors in the Canadian economy. Its successful adoption by industry will help strengthen our competitive advantage in the global market. However, we need to address the challenge of identifying relevant, viable, readily applicable AI-driven business solutions. We also need to identify the skilled individuals that businesses require in order to successfully adopt and implement AI technologies. Talent and skills should not be an afterthought in the adoption of AI; to ensure it is successfully implemented, this talent needs to be considered in tandem with the procurement of AI. The Brookfield Institute is working with partners to better understand and develop solutions to address the talent barriers preventing Canadian businesses from successfully adopting AI.
Championing ethics alongside AI adoption
Concerns related to physical, digital, and political safety as well as privacy, bias and discrimination are driving a growing number of conversations on the ethics of how AI technologies are developed and deployed. Just in December 2018, the G7 convened a multi-stakeholder conference on the responsible adoption of AI. In Canada, there has been a particularly high level of interest in these issues. The Treasury Board Secretariat is developing a directive on automated decision-making systems, along with an algorithmic impact assessment tool to help the federal government to deploy AI responsibly within its ranks. As well, the CIO Strategy Council is developing industry standards for data use and ethical AI.
Outside government, two conventions have emerged that promote the socially responsible development of AI: the Montréal Declaration for Responsible AI Development and the Toronto Declaration: Protecting the rights to equality and non-discrimination in machine learning systems. In an increasingly digital age, these are important steps toward ensuring that AI is developed in a way that respects human rights and promotes individual autonomy. Yet despite the growing community of researchers, organizations, publications and events focused on addressing ethics in AI, the landscapes in Canada and globally are heavily fragmented.
Canada could take a lead role in developing a set of ambitious and wide-reaching national guidelines on the responsible development, procurement, and implementation of AI technologies. However, guidelines on their own will not be sufficient. Our public service must be able to track technological developments, identify and conduct investigations into high risk AI applications, and report their findings publicly. Canada has the opportunity to set the stage for other actors to follow suit. Adding an ethical AI strategy to our already strong legacy of world-class AI research and talent would signal to our neighbours and competitors that we intend to lead in a thoughtful way.
Engage citizens on the value of their data
AI relies on data to function. In order to perform well the data needs to be accurate, but it also needs to be representative of the population it serves and the context in which it operates. One of the greatest challenges that companies face in deploying AI is acquiring the right quantity of high quality data. This challenge is inextricably linked to citizens’ perceptions; whether they have trepidation about giving companies access to their data.
In recent years there has been a shift in attitudes toward corporate data collection and privacy. There is a growing distrust of how personal data is protected and used, particularly regarding third-party access to personal information and browsing habits. It is perhaps best exemplified by the #DeleteFacebook movement that surfaced early in 2018. In this climate, companies have been forced to rethink their approaches to data privacy and security.
Many AI applications present opportunities to enhance peoples’ lives. More personalized or efficient services in industries such as health care and transport is one example. But these benefits will only be realized with adequate public buy-in and trust. Canada must chart a path through these concerns, to ensure the public benefits of AI can be realized without jeopardizing peoples’ privacy and security.
Canada can become one of the first nations to proactively engage its citizens in a meaningful dialogue on the benefits of making data accessible, protected and private. If we proactively launch these conversations, we could gain limitless opportunities to drive innovation, improve services and generate better outcomes for people. Together, we can set clear parameters for data use, ownership and governance to protect the public’s interest and maximize public benefits.
There will be huge social and economic gains if Canada makes it a priority to identify and develop the relevant business applications, strengthen talent pipelines to support AI industry adoption, create an ethical framework for the development and deployment of AI in the public and private sectors, and engage citizens in a discussion about and the design of data protection. In advancing these priorities, Canada will stand out as a global leader — not only in the development of AI, but also in its effective, thoughtful, and forward-looking implementation and governance.