With the novel coronavirus COVID-19 wreaking havoc around the world, AI researchers have been using machine learning to track the spread of the virus using social media, website and other data.
John Brownstein, an epidemiologist from Montreal, is the chief innovation officer at Boston Children’s Hospital and a Harvard professor, as well as part of an international team using machine learning to sift through social media posts, news reports, data from official public health channels and information supplied by doctors for warning signs of where the virus may be taking hold around the world.
The software being developed by Brownstein’s HealthMap AI team uses information gleaned from the internet “to come out with a more accurate prediction of how [COVID-19 and other viruses are] growing and where we’re going,” says Shahrzad Esmaili, a partner in the intellectual property group of Gowling WLG (Canada) LLP in Toronto.
It’s one way in which tech companies are harnessing artificial intelligence and how collaborative efforts are helping advance AI and other research. At the same time, companies must pay attention to the protection of their assets.
Canada as an AI and research hub
“AI is a very big sector” in Canada, says Isi Caulder, a partner in IP boutique firm Bereskin & Parr LLP in Toronto. “AI clients are coming in fast and furious,” she adds, with tremendous activity in Toronto and Montreal in particular. This runs the gamut from computer vision and medical diagnostics to machine learning and its subset natural language processing.
Canada has become a hub for AI, from Google’s smart Sidewalk Toronto project to the AI-Powered Supply Chains Supercluster (Scale AI) in Montreal, which focuses on artificial intelligence and supply chain technology. (The Scale AI supercluster is one of five in Canada that Innovation, Science and Economic Development Canada launched in 2017 and in which it will invest up to $950 million, to be matched dollar for dollar by the private sector.)
Canada’s superclusters are “wonderful experiments,” says Anthony de Fazekas, head of technology and innovation for Norton Rose Fulbright Canada LLP from his Toronto office. They allow common IP approaches to be developed, including templates, that put joint development deals together. “The next [innovation] will be companies collaborating on COVID-19 products.”
Collaborative agreements are “a great way to bring innovation to market in a creative way,” says de Fazekas. “We will see more of that consortium building happening within [the] technology scaleup part of the economy.”
Despite the COVID-19 pandemic, “parts of the economy are still moving forward, but there’s an increasing need to share products and intellectual property around things like ventilators,” he adds. “We’ve seen interesting projects solely based on open- and cross-licensing [and] open-data projects. It’s interesting to see . . . the intensification of open licensing and IP, even across companies that were holding their cards close to their chest.”
Today, there is an increasing urgency to not only bring certain innovations to market but, for a broad distribution of that technology, quickly, he says.
In academia, research labs are being established for cross-disciplinary collaboration, which is unique for this generation, says Gowlings partner Shahrzad Esmaili. She cites the Institute for Biomedical Engineering, Science and Technology, or iBEST, as a partnership between Ryerson University and St. Michael’s Hospital in Toronto that utilizes the research of various departments within Ryerson, such as biomedical and computer software.
“It’s a good example that I’ve seen,” Esmaili says, “and other universities are forming similar platforms for cross-research.”
Collaborative agreements between industry, academia and the public sector mean that technology and other companies must take special care to protect their IP.
Companies need to strike a balance between what they need to achieve and make openly available as part of their collaborations with a third party, says Panagiota Dafniotis, national lead of the intellectual property group at Dentons Canada LLP in Montreal. Particularly in environments for startups and collaborations within academic communities, who owns what needs to be established, as well as collaborative objectives.
“It’s extremely important to have great clarity regarding what joint ownership entitles you to do,” says de Fazekas. This includes defining what parties are hoping to build and recalibrating the commercial understanding around IP, trade secrets and more.
“Once that’s in place, you can have a much more robust and workable IP governance” regime for the collaboration agreement, he says. “If you start collaboration with clarity of a project plan, everything is more likely to fall into place.”
Even outside of collaborative environments, the protection of trade secrets and patents is fundamental to protecting a company’s intellectual property.
“I would love for more Canadian companies to consider AI protection as valuable and important to their companies,” says Esmaili. Companies based in other jurisdictions can be much more patent-savvy, she says, and she would like to see Canadian companies, especially technology companies, “more aggressive in patent filing” in key jurisdictions such as the U.S., Europe, China and India.
“If we want to be able to compete on a global stage, then we need to be more active in our patent-filing activity,” she says.
Tech companies also need to put processes in place to maintain their inventions as confidential or risk “losing their patent rights in most countries in the world,” says Caulder. If the invention has been disclosed publicly, it may no longer qualify as new, and patent rights in Europe would be lost. If the public disclosure was by the inventor, a one-year grace period may apply in countries such as Canada and the United States.
“But the best practice is to keep everything confidential until you’re ready to file for a patent,” says Caulder. Trade secrets should be kept confidential on a need-to-know basis, using encryption and locked cabinets. But not all innovations should be patented, she says, and companies should engage in “rigorous vetting” processes to determine what may be important to patent from a revenue standpoint, given the product’s technical features, and that budgets are finite and applying for patents is expensive.
Design protection in software is especially important, she notes, specifically what users see on a device when they’re shopping online. “When you think about the value of the ‘look’ of a store, such as the Apple store or Petro-Canada, the reality is now, more than ever, a user’s shopping experience is virtual and so IP needs to follow,” says Caulder. “So, protecting what [users are] seeing on the device is really important.”
Industrial design protection has exploded for software companies over the past few years, she adds, and “lots of companies are checking what they’re presenting to users” online.
On the horizon
So, where are IP and AI efforts being focused now?
“Definitely neural networks,” says Esmaili. This is a form of AI that can be used for speech and image recognition. “These are hot topics, with many of the voice personal assistants that you’re aware of,” such as Alexa and Siri.
Dafniotis, too, sees voice technology as being a “very big growth area for the next few years.”
“As customers have become more comfortable with voice technology, businesses of all kinds and all sizes have begun to develop many different voice applications,” she says. “Voice is really relevant to how you serve your customers [and creates] a better client experience, with less friction. I think there’s greater market penetration opportunities [and] a very robust area for innovation, for understanding and protecting the advancements in this space.”
Voice devices that use automatic speech recognition and natural language understanding and are enabled in a hands-free environment are of particular interest and relevance in the COVID-19-era of physical distancing and non-touching. But, generally, says Dafniotis, “I think that we’re seeing clients and customers want to interact in a hands-free activated world.”
One AI development that still needs to be approved by regulators is “AI as inventor” — where there is no natural inventor of an invention but rather the inventor is a machine. One such artificial inventor is DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), a patented AI system developed by scientist and academic Stephen Thaler, that invented a container. Both the European and U.K. patent offices refused to grant a patent for the container, says Caulder, because there was no natural inventor.
“The designation of an inventor is mandatory as it bears a series of legal consequences, notably to ensure that the designated inventor is the legitimate one and that he or she can benefit from rights linked to this status. To exercise these rights, the inventor must have a legal personality that AI systems or machines do not enjoy,” the European Patent Office (EPO) explained in its reasons, released in January.
However, Caulder says, the director of the U.S. Patent Office, Andrei Iancu, has suggested that artificial inventors should be recognized for patents, and the decision of the EPO could be appealed.
In Canada, Caulder is hopeful that the Canadian Intellectual Property Office will soon “shine the spotlight specifically on AI technologies and give a bit more guidance. . . on how to how to help our clients protect inventions in that area.”