By Garth Gibson
Oct 26, 2022

For countries, succeeding in the knowledge-based economy is about more than education and innovation. Increasingly, it’s about securing and utilizing a nation’s intellectual property (IP).

That’s especially true in the era of Artificial Intelligence (AI), which holds the potential to improve decision-making, reshape industries, and accelerate growth. 

IP laws exist in part to encourage ingenuity and reward risk-taking. They offer a range of protections designed to shield an invention or creation from being easily copied or closely mimicked. But these laws are only effective when businesses understand them and take practical steps required to protect their IP.

Around the world, economic growth and productivity are increasingly being fueled by the so-called intangibles economy – things like data, services, brands, and licences. Our future prosperity is linked to our ability to thrive in these intangible spaces. AI is emerging as a key element of the intangibles economy. Indeed, as the Public Policy Forum has put it, “Technology is creating a whole new set of determinants of national wealth.” Make no mistake: in today’s knowledge economy, intellectual property is king.

At the Vector Institute, from the beginning, we have placed a priority on educating business leaders, innovators, and researchers about the importance not only of investing in AI – but, where possible, of protecting that AI innovation as IP.

Admittedly, this can be a challenge. The longstanding and pre-eminent way of protecting a breakthrough product is via the patent system. This is an important step, but it is not the whole story, as some aspects of AI development, such as mathematical formulas, can’t be patented. Put another way: You can patent an application that uses a formula, but not the formula itself. Additionally, the very act of submitting a patent for an AI model could conceivably provide competitors with sufficient information to cobble together their own version of the innovation. 

It sometimes can be situationally useful or even essential for companies to consider a multi-pronged approach: patenting the business application of an AI model while protecting the underlying equations by invoking and enforcing confidentiality agreements that help ensure that tech innovations stay with the company where they were conceived and developed. In practical terms, this can mean that only a device’s non-AI parts are patented – even if what makes it highly effective is the computer processing inside the device. Companies may also be able to use copyright to protect their innovation – the idea being that the trained model embedded in a software application can be licensed (and therefore both protected and monetized) as a software product. As an aside, there is already a growing public policy debate, driven by surprisingly satisfying artistic compositions generated by AI using as little as a few words and the commercialization possibilities of NFTs, over how much “human input” is required for an innovation to qualify for copyrighting.

Companies that fail to protect and secure their IP run the risk of having their ideas and innovations copied or stolen. In today’s highly competitive marketplace, that can mean the difference between being a leader or an also-ran in an industry. Good IP practices can help to mitigate this risk. For example, the benefits of being one of Vector’s industry sponsors do not include any rights or entitlements to Vector-created IP. In all of Vector’s sponsorship agreements, any IP created at Vector belongs to Vector.

In Ontario, the government recently announced the launch of Intellectual Property Ontario (IPON), the first agency of its kind geared to helping the province’s businesses and researchers protect and make the most of their IP, whether it is through patents or other strategies. This is a valuable initiative that highlights the growing focus on IP, its current and future value, and its singular role in building competitive modern companies and fueling modern economies.

At Vector, given our focus on deep learning and machine learning – we encourage businesses to incorporate AI in their processes and products. This is accompanied by a focus on also encouraging businesses to adopt the rigor required to build a strong AI IP portfolio.

This kind of AI/IP guidance is a key element of our work. We frequently invite our Vector partners together to encourage collective exploration of the latest AI methodologies. Our partners then take these insights back to their companies and labs, and with their own data and business expertise, invent, protect and apply their IP in the specific context in which they operate.

Protecting IP becomes even more nuanced in the context of research, where sharing ideas, and sometimes data and methods, can be essential to progress, innovation, modernizing regulation, and demonstrating responsible development of AI. Additionally, researchers can sometimes be cross-appointed or employed by multiple organizations at the same time. They may release copyrighted code or file patents independently, or through one or more employers, or in partnership with an institution such as Vector. Working with our university partners, our goal is to assist AI researchers in adhering to the so-called “scientific hygiene” required to ensure that any IP they invent is appropriately protected.  

Indeed, one of the most effective ways that Vector helps researchers become potential entrepreneurs is by helping to identify who invented what and when, making it harder for a predatory actor to take advantage of a nascent business. 

Finding the right ingredients for a successful AI model demands creativity, determination, and patience. But that’s only the beginning. Protecting that model – and the rest of a company’s IP – is equally important. The successful companies of tomorrow will be those who develop the vigilance and diligence required to shield and secure their innovations.

If you’re a startup/ scale-up and you want access to Improve your commercialization and IP strategies, sign up for Vector’s FastLane here.