In my previous post I argued that if Canada wants to succeed with its AI-focused innovation agenda, it should also be at the forefront of 5G innovation and development. Canada could get ahead in the global 5G race not by being the first to 5G, but by being the first to roll out 5G in the right way – addressing cybersecurity, linking development of AI and 5G, addressing regulatory and policy prerequisites, etc.

Canada could leap ahead in development of both, 5G and AI, by tackling them collaboratively rather than in parallel but separately as is the case now. One idea comes from the U.S.

5G Innovation Zones in the U.S.

In September 2019, the U.S. Federal Communications Commission (FCC) announced its creation of Innovation Zones in designated portions of New York City and Salt Lake City, Utah, where experimental licenses will be granted to allow testing of 5G – beyond what can be accomplished on a university campus or in a corporate lab setting alone. This initiative comes in response to a proposal from the Platforms for Advanced Wireless Research (PAWR) program, which anticipates “exploration of robust new wireless devices, communication techniques, networks, systems and services that will revolutionize the nation’s wireless ecosystem, thereby enhancing broadband connectivity, leveraging the emerging Internet of Things, and sustaining… economic competitiveness for decades to come.”

The New York City innovation zone will support COSMOS (Cloud Enhanced Open Software Defined Mobile Wireless Testbed for City-Scale Development) in West Harlem. It will be operated jointly by Columbia University, Rutgers University and NYU. Meanwhile, AT&T is deploying AI in software-defined networks (SDN) more generally and using network function virtualization (NFV) with a goal of 75% virtualization of its core network by next year. The combination of SDN and AI will enable the high speeds and low latency that is the promise of 5G.

Salt Lake City’s innovation zone will support POWDER (Platform for Open Wireless Data-driven Experimental Research with Massive MIMO Capabilities) and run “connected corridors” with multiplied capacity, overseen by Rice University and the University of Utah.  AI will enable 5G ‘networking slicing’ offerings to customers that will enable multiple applications to run simultaneously over the same connection.

AI and 5G Innovation

Independently, the potentials of AI and 5G innovations are receiving mountains of attention and investment. But it is the marriage of these technologies, as well as the Massive Internet of Things (mIoT), that promises the sea change so many have been expecting from the technological revolution. In an almost ‘chicken and egg’ scenario, 5G is needed to enable use of AI at scale and drive mIoT use cases, and demand for AI and mIoT is needed to build the business case for the huge investment 5G networks require. The parts are powerful, but 5G innovation, AI and IoT together create a meta-system capable of realizing previously unknown possibilities.

You can read more about the opportunities, and risks, of tackling 5G and AI together in my previous articles: Digital Double Helix: Why the Fates of 5G and AI are Intertwined, AI at the 5G Core: A Double-Edged Sword and Getting Smart About 5G Networks and AI in Canada.

There are few more reason to tackle two technologies jointly.

Leveraging AI for Deploying 5G

Separately from government-facilitated innovation zones in U.S., AT&T is testing the use of AI in the form of drones with video cameras flying around cell sites in New Jersey testing radio transmitter signals.  They also test whether a cell is functional, or whether rust, dirt or disconnected wires are causing a malfunction. The drones can determine remotely whether humans are required to go and climb the cell tower at any particular site for repairs.

If AI detects a cell site isn’t functioning properly, it will signal another tower to pick up the slack. If one area is experiencing a spike in usage, AI will trigger lower-use cell sites to maintain throughput.  AI can be used for network routing of high volume 5G data traffic, such as real time video.  It can even detect or predict congestion on the small cells of 5G networks before it happens. By putting intelligence closer to the edge, telecoms are starting to load balance traffic across small cells as will be necessary with commercial 5G.

The biggest cost for all telecom companies related to 5G is in planning, designing, and installing the network, then maintaining and securing it. AI and machine learning (ML) can be leveraged to map cell towers, fiber lines, and other network transmitters and to help figure out the best locations for build out of macro cells, smaller cells and picocells.

As mentioned, in order to benefit from 5G innovation a 5G network needs AI as a key ingredient for making it autonomous, and able to diagnose and repair itself at its mobile edges and local radio access network (RAN). Telecoms are also using AI already to manage their on premises cloud, hybrid cloud and third party cloud computing. They’re pushing some of that intelligence both to the local RAN and to national, local and regional data centers.

AI in Canada

Canada lands on the shortlist of global leaders in artificial intelligence (AI) thanks to its five bustling technology hubs that are well supported by academia and industry with a nice helping of government funding. The AI innovation hubs are found in Edmonton, Montreal, Toronto, Vancouver, and Waterloo, with the nation’s capital, Ottawa also getting into the game. These superclusters were tied together in 2017 by the federal government when it launched a $125-million Pan-Canadian Artificial Intelligence Strategy, joining the U.S., China and France in adopting a national plan for AI.

In September 2019, the governments of Canada and Quebec announced the formation of an international center in Montreal for the responsible advancement and adoption of AI, as part of the Global Partnership on AI, recently discussed at the G7 summit in August in France. This brings further recognition to Montreal’s tech ecosystem as a leader in AI expertise. The center will bring together public, private and academic stakeholders and focus on the ethical, human rights, inclusion, and diversity issues raised by emerging AI technology, as well as its commercial innovation and economic aspects. It will assist the work of Canada’s Advisory Council on Artificial Intelligence.

Alongside enviable degree programs in AI, The University of Alberta boasts the Alberta Machine Intelligence Institute (AMII) program in Edmonton; the University of Toronto has the Vector Institute and the University of Montreal and McGill now share affiliations with the Montreal Institute of Learning Algorithms (MILA), complementing the work of Yoshua Bengio’s Element.AI.

The Toronto-Waterloo corridor, dubbed ‘Silicon Valley North,’ is a frontrunner in AI and the second largest tech region on the continent. Vancouver on the other hand, is home to an Amazon/AWS expansion focused on cloud computing and machine learning along with AI start-ups like Mobify, Sancuary.AI, Finn.ai and a team of AI researchers from the University of British Columbia. Ottawa, as the city with the most highly educated population in Canada, boasts more engineers, scientists and PhDs than any other. AI companies run on data and the Canadian government generates plenty of valuable big data that can be leveraged most readily by AI startups in Ottawa with expertise navigating the government, whether in cybersecurity or other fields. Further, the telecom regulator with jurisdiction over 5G networks, the CRTC, is based in Ottawa. BluWave and Cognos Analytics, which is now a division of IBM, focus on business intelligence in Ottawa.

All of this advanced technology being developed in Canada will need superfast, low latency connectivity if it is to be successfully commercialized and scaled up for use by Canadian businesses and implemented by national, provincial and municipal governments in support of public services. The technology for that connectivity is available in the form of 5G next generation wireless, but the actual physical network infrastructure for that service is mostly lacking so far, as it is in the U.S. Telecommunications companies want to see real world demand for 5G connectivity and the Internet of Things (IoT) from enterprise customers to justify investing more of the large sums needed to build out 5G networks. Alternatively, they would want to see 5G innovation that would reduce upfront infrastructure costs.

The Challenge of Commercializing AI in Canada

One major stumbling block to commercialization is that a big chunk of the intellectual property, mainly patents, that are developed in Canada don’t stay in Canada, but are sold to foreign corporations. Thus, Canadian AI tech innovation hubs are sometimes thought of as “branch plants” for foreign projects like helping Hollywood write better scripts and helping big tech company partners with data anaytics. Apparently fewer than half of the machine learning patents developed in Canada between 2007 and 2017 stayed in Canada, according to a think tank founded by the former co-CEO of Blackberry. Sometimes local startups end up having to license those patents back at steep rates to implement those same innovations in their own businesses.

On the academia front, the University of Waterloo’s Engineering department has launched Robohub, which houses Talos, its humanoid robot. Talos represents one of the global frontiers of machine learning research, but critics disparage the generous funding of such research at the expense of commercialization.

5G Innovation in Canada

Rogers has invested $4.7B in 5G in 2019 alone, working with Swedish equipment supplier Ericsson. Further, it has launched a second 5G innovation hub and testbed in Ontario in collaboration with the University of British Columbia to help Canadian companies and researchers use 5G in optimal ways. Rogers’ first partnership with UBC on the Vancouver campus concentrates on research, while the lab at Communitech in Waterloo, ONT will advance “made in Canada” 5G innovations and technology and focus on commercialization of IoT systems and smart cities.  The company recognizes that it needs to build a complete “ecosystem with the right partners, spectrum, infrastructure and investments to make 5G a reality.” Communitech is public private partnership that has already worked on scalable 5G innovation and business use cases with well over a thousand companies ranging from startups to legacy enterprises.

Rogers recently announced a reciprocal roaming agreement with AT&T to extend IoT coverage for enterprise customers of both companies throughout Canada and the United States.  This will benefit businesses that use near real-time IoT connectivity solutions, such as freight service companies tracking vehicles and any companies that use telematics for improved workplace safety.

Telus Mobility is testing 5G in Vancouver where its Living Lab, originally announced by Huawei, aims to make Vancouver “the greenest city” by 2024. Given Canada’s expansive geography, the benefits of 5G may be mostly local in many places where MIMO can connect thousands of fixed and mobile sensors and devices to a small, dedicated network such as a civic IoT for a smart city, or an industrial IoT that is then linked via 4G backhaul to a regional data center and the core network. Sometime soon, the many millions of consumer sim card connections in a city will no longer comprise the majority of telecom network endpoints.

AI + 5G Innovation

Advanced AI research and development clearly is thriving in Canada, and its telecom companies are working on developing their 5G networks, but most of this activity is happening in parallel rather than collaboratively.

Perhaps it would make sense for leaders in each of the five AI superclusters in Canada to work together with top executives of its big telecommunications companies to create hybrid AI+5G innovation zones and boost this next wave of digital technology transformation that promises healthy dividends for economic growth.

For example, above mentioned Rogers’ partnerships with UBC in Vancouver and Waterloo could collaborate with ongoing AI hub activity in those cities to help develop business use cases for AI that will establish more tangible commercial demand for 5G connectivity.

By linking use cases for the two cutting-edge technologies, organic Canadian AI+5G innovation zones could result.

In terms of spectrum allocation for 5G, Canada, in contrast to Europe, has the potential advantage, even with its larger geography, of being able to adopt one technical standard and one unified frequency allocation plan. However, Rogers and Bell Canada have much of the prime 3.5 GHz band for example, already licensed instead for fixed wireless deployment in rural areas, so some re-auctioning may be in the offing. Regional spectrum auctions for 5G could form an early action item for innovation zones in Canada.

Joint AI+5G innovation zones would also create better commercial opportunities where the benefits of them together would be greater than the sum of its parts. This could drive more VC and government funding to balance the investment from U.S. companies and would help get more AI or 5G applications commercialized inside high-growth Canadian companies. Addressing the AI commercialization challenges mentioned previously.

marin@5g.security | Website | Other articles

Marin Ivezic is a Partner at PwC (PricewaterhouseCoppers) specializing in risks of emerging technologies. He leads PwC’s global 5G cybersecurity efforts. He also leads cybersecurity for the Telecommunications, Media & Technology sector; and Industrial, IoT, Critical Infrastructure & Cyber-Kinetic security capabilities in the region. All these focus areas are being transformed with the emergence of 5G. Marin worked with critical infrastructure protection organizations in a dozen countries, 20+ of the top 100 telecom companies, and a number of technology companies on understanding the geopolitics of 5G; uncovering as-yet-unknown security and privacy risks of 5G, AI and IoT; and defining novel security and privacy approaches to address emerging technology risks.