Anthony de Fazekas is head of technology and innovation in Canada and an intellectual-property partner and patent agent, and Maya Medeiros is an intellectual-property partner, patent and trademark agent, at Norton Rose Fulbright.

Canada, a world leader in artificial intelligence and deep learning, is home to research pioneers, and its talent now attracts global companies to come here and set up research labs. With other countries such as China accelerating the pace, the race is now on to produce innovative AI-enabled products and services.

To compete we must be strategic in how we marshal our resources, and a key factor will be intellectual-property management, particularly in collaborative AI developed by different stakeholders. Determining who owns or controls the IP rights of a new technology and, in turn, who will be rewarded for their expertise and efforts, can get messy if contracts do not exist or do not contain clear IP terms.

Story continues below advertisement

If Canada is to be a true leader in AI, it is imperative that the legal regime that underpins it provides clear protection for AI innovations and Canadian data that are flowing out of the country. Currently, IP law does not provide a clear answer to key questions: Does a party contributing data really own the data? Who owns the AI algorithms? What other data does the data set depend on and who owns those other data? What rights of use apply to those other data? Is there clear chain-of-title across all the sources of data to the final data set?

Canadian companies, on average, spend less on IP protection than their foreign counterparts and own less IP. Valuable AI solutions that are built by Canadian companies can fall squarely within the patent rights of competitors who can enforce these rights to prevent freedom to operate and extract patent royalties. Developing valuable patents may be expensive and time-consuming, but should be prioritized to protect the freedom to operate and commercialize AI innovations.

The federal government’s $950-million superclusters initiative is an example of a collaborative framework essential for innovation that will generate jobs and wealth in Canada from AI, and it can also be part of the IP and data solution. By providing frameworks that will enshrine practical mechanisms for devising agreements sharing the AI value chain, the superclusters will help to resolve ownership of AI and data.

The joint projects across startups/scale-ups, large companies and universities – which can often be based on unexpected synergies – not only promise to generate significant commercial value, but also to retain ownership data and AI innovation. An emerging business might develop a leading AI technology but requires industry-specific data to develop a solution. An established company in that sector will have industry-specific data. The potential for creating value if these two companies come together and share resources is enormous.

Legal reform may also have a role to play in AI, particularly around data. The British Columbia Freedom of Information and Protection of Privacy Act (FIPPA) represents a useful model to consider in relation to AI. It requires that public bodies store any personal information that is under their control or custody in Canada only, making it only accessible in Canada, except for a few exceptions. This has resulted in a treasure trove of health data that give B.C. a competitive advantage in building digital health companies and attracting foreign tech companies to operate there.

A similar approach can be taken in a variety of sectors in Canada, such as agriculture, energy, infrastructure and natural resources. One way would be to require that data generated in Canada at least be stored in a Canadian data warehouse, while also being permitted to leave our borders. This may ensure the buildup of significant Canadian data assets while also avoiding substantial restrictions on multinationals operating here, thereby making Canada a competitive jurisdiction for both locals and multinationals to run corporate innovation projects.

By bringing together executive and technical talent, resource commitments, energy and determination from teams to drive commercialization forward, the superclusters bode well for the future of innovation in Canada. Our AI community should align with the superclusters, learn from them and participate in them. Yet, it remains that a clear IP framework for AI innovation is needed in Canada.