Artificial Intelligence, Real Impact: How Canada can leverage domestic excellence in AI
While the Canadian private sector has been quick to adopt home-grown Artificial Intelligence-enabled products, these capabilities have been adopted more slowly in government. There is a unique opportunity for Canada, in partnership with industry, to improve public services and optimize government functions – including national defence and domestic security – by leveraging home-grown AI capabilities.
Senior Analyst Alessandra Jenkins explores Canada’s current AI innovation infrastructure, challenges for industry engagement with government, and opportunities for the Canadian government to leverage AI across a variety of functions and services.
Canada’s researchers and industry are at the forefront of artificial intelligence, a capability in demand globally in a number of diverse sectors. Canada is home to some of the world’s top AI innovators.
The country houses multiple hubs of AI excellence, attracts investment from leading international companies, and has government funding to sustain research. Nation states, companies, and researchers worldwide turn to Canadian talent to develop sophisticated algorithms and AI-enabled technologies to serve a wide range of functions.
While the Canadian private sector has been quick to adopt AI-enabled products, these capabilities have been adopted more slowly in government. There is a unique opportunity for Canada, in partnership with industry, to improve public services and optimize government functions – including national defence and domestic security – by leveraging home-grown AI capabilities.
This paper will explore the state of Canada’s AI innovation infrastructure, challenges for industry engagement with government, and opportunities for the Canadian government to leverage domestic AI talent across a variety of functions and services.
As artificial intelligence becomes more prominent and popular, the term “AI” is often used interchangeably – and imprecisely – to describe a number of different computer processes. At its root, Artificial Intelligence is an area of computer science that uses algorithms to create intelligent machines.
These algorithms, rather than humans, allow a machine to take information and react to it. Machine learning, which is part of AI, is the ability for the computer to learn on its own without subsequent human input or supervision.
Machines that have access to an abundance of information are programmed to identify patterns, classify information, and provide outputs in similar ways that humans do for problem-solving.
The ability to recognize speech, plan, learn, and problem solve is what makes artificial intelligence unique. This ability can be layered with other technological advances, such as robotics, to create new capabilities such as unmanned systems.
Developing a strong understanding of the various technologies that enable AI and their various applications is an important first step to understanding how it can best be applied in various government applications.
Canada’s AI Leadership
Canada’s AI capabilities have been nurtured by public initiatives, growing private investment, and strong research from industry and academia. The result is an AI ecosystem that spans provinces and institutions, positioning Canada as a global leader.
The World Economic Forum indicates Canada has the fifth largest AI workforce globally, while Times Higher Education ranks Waterloo as one of the top 15 universities for AI in the world.
The Canadian government has demonstrated financial support across the country for AI research. In 2017, the federal budget enhanced funding to the Canadian Institute for Advanced Research (CIFAR) to launch the Pan-Canadian Artificial Intelligence Strategy.
This $125M investment promotes collaboration, graduate training, and allocates CIFAR AI Science Chairs across centres of AI excellence including Toronto-Waterloo, Montreal, and Edmonton.
Other government programs have also supported AI, with SCALE.AI being recently selected as one of Canada’s Superclusters. The joint Quebec-Ontario cluster aims to transform industries by building intelligent supply chains, a project that anticipates creating 16,000 new jobs and adding $16.5 billion to Canada’s economy over the next decade.
Additionally, large technology companies such as Google, Facebook, and Uber have committed funds and labs in Canada dedicated to AI. Facebook alone invested $7 million into Montreal’s AI community, to include launching a new lab, and plans to double the number of AI research staff.
Facebook has also seen early successes with an AI algorithm project for MRI machines with a US university. These and other investments demonstrate that Canadian AI capabilities are being embraced by global commercial markets. However, adoption of AI in the Canada’s public sector is proceeding at a slower pace.
AI for Public Sector
There are a number of opportunities to leverage Canadian-sourced AI to optimize key government functions and improve services. The plethora of domestic capability can provide Canada with needed capabilities while the government can provide an end-user customer to drive the development of public-sector AI applications. The major areas in which we see near term government adoption of AI is as follows:
Data Management: Processing, disseminating, and analyzing vast amounts of data across government departments including tax, healthcare, and multilingual citizen services.
Cybersecurity: Leveraging advanced AI to better detect, analyze, and mitigate the persistent barrage of cyber threats faced by government networks.
Autonomous Systems: Using unmanned systems to improve key government functions and services, including defence, homeland security, and research.
Leveraging data to its maximum potential is needed in order to improve decision making that will direct Canadian government services and policies. Machine learning and language processing functions have potential to support government in managing data and disseminating information. Employing algorithms and applications that can interpret demographic and
behavior data provides numerous benefits to both public servants and Canadian citizens consuming government services. Successful use of AI by foreign governments and commercial industry demonstrates how these capabilities can be accessible and applicable to Canada.
CRA and Tax Collection
AI can improve the tax filing process for citizens. Prior to the April 2018 tax filing deadline, the Canadian Revenue Agency (CRA) reported that over 90% of the nearly 20 million returns were being filed electronically. AI can supplement the efforts already underway at CRA (e.g., Interactive Voice Response System) to most efficiently disseminate information to Canadian tax payers.
Tax agencies and companies in other countries already use AI algorithms to store and recall tax codes. This can augment civil servants’ efforts to publicize accurate tax information coast-to-coast-to-coast.
For example, the Australian Taxation Office website launched a virtual assistant, ‘Alex’ in 2015 to support personal tax questions. The virtual assistant uses natural language processing to understand conversational language and provide relevant answers to questions from Australian taxpayers.
For CRA civil servants, AI can identify patterns in records at greater speeds than humans by clustering filings from companies or citizens with similar tax characteristics. Using AI for analytics can directly support the public service in making evidence-based decisions by providing comprehensive data at faster rates.
Additionally, AI is being leveraged in the private sector to assist with simplifying tax filing. H&R Block presently uses IBM Watson within their financial solution software, which enables accountants to quickly comb through thousands of pages of tax code to find solutions for customers.
This allows accountants to spend less time doing tedious administrative tasks and more time engaging with customers – a development that benefits both tax filers and professionals.
Delivery of Healthcare Services
The combination of an aging Canadian demographic and demand for physicians, which is growing at twice the pace of the population, is putting strain on health care providers and on government budgets. Provinces are well aligned to become end-users of AI, with some already adopting AI in pilot projects to manage patient records.
Over their lifetime one person creates the equivalent of 300 million books in health-related data. Hamilton Health Services currently has a project underway to use cognitive computing to take in data, learn from it, and provide suggestions.
This type of support provides for easier access to information, minimizes time researching, and could allow doctors to both take on more patients and spend greater time on person-to-person care. Further opportunities exist in this space for industry and government to collaborate.
For example, projects such as the previously mentioned MRI pilot conducted by Facebook in Montreal could maximize usage of machines to service more patients.
Inclusive, Multilingual Government Services
The 2011 census indicated over 200 languages are spoken across Canada with 6.8 million people reporting a mother tongue different from the two official languages. While many people are multi-lingual, expansion of services across languages can ensure all Canadians have full access to government services.
While chatbots have been used for government services like the taxes, the changes in technology are already creating opportunities to broaden the scope of their use. Advanced machine-learning algorithms are estimated to have the ability to capture 80-90% of speech, making chatbots highly effective.
Most chatbots are currently developed in English, Chinese, Japanese, or Korean to meet commercial market demand but there is the emerging capability to develop products that are multilingual. Canada can explore acquiring chatbots that can deliver services in both the two official languages and for other commonly used languages in the country’s regions and communities.
For example, Immigration, Refugees and Citizenship Canada can leverage chatbots to guide Canadians on passport applications and renewals or direct visitors and potential immigrants to the correct processes for their needs. By becoming a customer Canada can provide an end-user for companies developing these emerging capabilities while increasing accessibility of government services.
Understanding the Government’s Cybersecurity Challenge
Protecting sensitive citizen and national security data has become a critical responsibility and area of investment for government. Threats to Canada’s cybersecurity span terrorist organizations, nation states, ransom hackers, and unintentional “insider threats;” all of which have different mechanisms and motives for obtaining access to government held data.
The threat is persistent and can target a range of government functions: from large-scale international operations, such as the hacking of 4.2 million personnel files from the Office of Personnel Management in the US in 2015, to domestic incidents such as the recent breach of tens of thousands of patient records through an Ontario health care service, CarePartners.
In 2017, Communications Security Establishment (CSE) released a report  that stated the Canadian government is attacked by a state actor nearly 50 times a week; at least one of those attacks, on average, is successful.
Government organizations including the Department of National Defence (DND), Public Services & Procurement Canada (PSPC), and the Communications Security Establishment (CSE) will be responsible for investing in and implementing the next generation of AI-enabled cyber defensive tools.
Canada’s 2018 budget reflects cyber capabilities are of central importance to the government – a $500M increase over five years – to include investment in enabling technology and standing up the Canadian Centre for Cyber Security under CSE. AI will increasingly become the essential equalizer needed to navigate and defend against the pace and persistence of cyberattacks.
AI-Powered Cyber Defence
Machine learning will continue to fundamentally transform cyberspace – impacting everything from sorting emails, to creating powerful and sophisticated cyber threats and responsive defensive mechanisms.
Advanced AI algorithms will be able to automatically detect vulnerabilities in government networks and launch attacks which are persistent, targeted, and nearly impossible to combat with human cyber analysts or traditional protocols. Cyber threats are constantly evolving and growing more advanced. Defences need to be as adaptive and sophisticated.
AI enables rapid and actionable analytics on network traffic that allows IT technicians to have a full understanding of the types of, and severity of, threats hitting their network and invest and prioritize accordingly.
Additionally, certain tools will be able to automatically respond via a pre-set playbook to common threats, freeing up the time of human cyber defenders to focus on higher-impact issues. Leading cyber companies and boutique AI-focused companies are leveraging AI to deliver cutting edge security capability.
For example, Symantec’s Endpoint Detection and Response software detects exploits and breaches and then deploys on-demand responses to recognized threats.
Leveraging Home-Grown Capability
Canada’s robust financial sector has created a hotbed of tech talent focused on supporting the sophisticated networking, computing, and security needs of the leading financial institutions that covered Toronto and beyond. This IT legacy paired with access to next-generation AI technology has spawned research centres of excellence that focus on AI-enabled cyber capabilities (e.g., Vector Institute, Element AI) and industry leaders (e.g., Phirelight, Maluuba).
Global tech companies including IBM, Dell, CISCO, and Thales have worked closely with Canadian researchers to mature AI technologies and grow the prevalence of AI-enabled security systems. Canada is home to world-leading full-service security companies such as CGI, which recently received a $530M US government contract for cybersecurity support over a large set of US-owned companies.
Agencies including PSPC, CSE, and DND must prioritize collaboration with private sector talent to develop sophisticated, custom solutions that can defend and secure systems and data across the government enterprise.
Engaging more UAS can help alleviate the strain placed on limited pilot resources by Canada’s vast geography and far-flung borders.
From driverless cars, to UAS, to undersea unmanned vehicles, AI is a fundamental technology driving autonomous systems. Government can employ AI to help overcome the challenges posed by Canada’s vast geography and high costs of traditional infrastructure.
Acquiring autonomous systems to meet research, operational, or service needs can create opportunities for higher quality of life forCanadians while supporting development of new capabilities in industry.
Defence And Domestic Security
Canadian Forces have already begun to adopt Unmanned Aerial Systems (UAS) technology, with the Army receiving a Blackjack UAS system from the US in 2017. The DND should consider supporting pilot programs to leverage fully UAS – in conjunction with AI where appropriate – for military missions.
For example, the US Defense Department presently has a program exploring how large UAS can be used to transfer cargo, minimizing the threat to soldiers who traditionally pilot an aircraft or drive to get supplies to service members in theatre.
Whether its providing greater surveillance and intelligence information, or minimizing exposure to threats for logistical tasks, AI-enabled UAS can provide rich data analysis to improve the safety and security of Canadian service members.
Bolstering the use of UAS for surveillance is also helpful for domestic security. For example, engaging more UAS can help alleviate the strain placed on limited pilot resources by Canada’s vast geography and far-flung borders.
Arctic Research And Service Delivery
Ongoing initiatives provide opportunity for industry to pilot new UAS capabilities. The Arctic is nearly 40% of Canada’s land mass but is home to only a very small percentage of the population.
Threats on Canadian Arctic sovereignty are growing from both from traditional countries like Russia and new countries like China vying for resources and trade routes as climate change creates new opportunities. AI-enabled UAS can be leveraged to enhance manned surveillance in difficult climates and improve livelihoods of communities in the long term.
Transport Canada already leverages autonomous vehicles and can continue to employ advanced technologies to provide high quality service to Arctic communities with limited roads and delivery options. The Arctic Unmanned Aircraft System Initiative is conducting drone trials as they decide on purchasing UAS.
Using the Sea Hunter Drone, the government is looking to use drones to detect oil spills, survey ice and marine habitats, and monitor activities. As UAS become normalized and beyond-line-of-sight is more common, Transport Canada can encourage delivery of essential goods and services by UAS. This can reduce the costs of goods for remote communities over time and improve purchasing power for families in the Arctic.
Canada can also use autonomous vehicles to support environmental research initiatives. Non-profits such as the Alaska Whale Foundation and the Alaska Ocean Acidification Network use autonomous systems to monitor wildlife and sea ice.
Using Intel Xeon and Movidious technologies, they collect water samples from whale blowholes to track their health in real-time, allowing them to understand current conditions for wildlife. For sea ice, Autonomous Surface Vehicles are used to assess the continental slope for the Chukchi Sea and Canada Basin to understand the changing ecology of the Arctic Ocean.
With vast spaces to cover and climate change concerns, technologies can support research initiatives from the Department of Fisheries to Environment Canada and equip researchers with the tools to monitor and report on the environment in real-time.
How can the government prepare to cultivate and integrate the best of Canadian technology into their future AI-based systems and processes?
- Develop a refined understanding of core AI technologies, the breadth of the Canadian industry and R&D, and scale of potential AI use cases for government.
The term “AI” has become increasingly comprehensive, ambiguous – and misunderstood. Understanding the specific technologies that comprise “AI” can be a challenge. Additionally, it can be difficult to keep track of new AI developments around the globe as well as the expanding number of Canadian AI-related companies (established and startups) and subject matter experts. The government should develop a better understanding of the potential of AI applications beyond those highlighted in this paper.
- Identify and organize key government influencers and decision makers who will be responsible for the facilitating collaboration with industry and responsible for integrating AI into government processes.
Creating an effective change management ecosystem can be essential to ushering in meaningful government adoption of AI. The right people must be in place to collaborate across need identification, sourcing, consultation, and implementation of this technology. For example, collaboration between the DND and Innovation, Science, and Economic Development Canada could create more opportunities to match domestic AI capabilities to defence-specific needs.
- Set metrics and goals to measure progress and effectiveness of AI integration into government applications.
Identifying objectives and timelines can be essential to measuring progress. Both small goals (such as meet with three start ups per quarter) and larger goals (increase speed of three core software based processes with AI) can help drive and focus efforts towards demonstrable AI achievements. Consulting with technological and economic experts can help officials craft realistic metrics for helping the government as it continues to mature technologically with the help of AI.
How can industry help answer the call?
With the proliferation of start-ups, SMEs, and large, diversified tech companies investing in AI-related technology, it is unlikely that the government can retain an up-to-date repository of all companies and their offerings.
Canadian companies must clearly articulate use cases and the right market price in a convincing and clear manner.
Companies with public sector experience as well as commercially-focused businesses need to proactively engage with government stakeholders through trade organizations, universities, and scanning open-source publications.
The government can offer attractive R&D opportunities, as well as a beta customer for emerging concepts and mature products, providing new revenue streams for a range of companies across the industrial base.