On February 21, 2019, Blackberry completed its acquisition of Cylance, a privately-held artificial intelligence (AI) and cybersecurity company. Acquisitions of AI companies like Cylance are becoming increasingly common as businesses seek to realize the opportunities in offering much-improved products or services to their customers. Canada, in particular, has become a hotspot for activity in the AI industry.

Acquiring an AI company is not always smooth sailing. There are common risks that buyers must be aware of prior to embarking on an acquisition.

Know where the data comes from

An AI derives its value from data sets used to train the AI. A common phrase in the AI industry is “garbage in, garbage out” – meaning, an AI is only as good as the data that has been used in its training. Canada regulates the commercial use of personal information through the Personal Information Protection and Electronic Documents Act (PIPEDA). PIPEDA requires consent of individuals before a company collects, uses or discloses their personal information during commercial activities.

Where the AI has been trained using personal information or uses this data in the course of its operation, PIPEDA requirements will be triggered. As such, during the due diligence process, the buyer must be wary of the consents the target has received from its customers. Often, these will be included in the standard terms of use on which the target engages its customers.

IP owner-“ship”

As with purchasing any technology company, reviewing intellectual property (IP) ownership will be a key part of the due diligence process. Often the sole value of the AI company will be derived from its IP.

In connection with traditional IP ownership issues, buyers must be cognizant that many companies in the AI space are not actually creating their own free standing AI systems but are instead leveraging those created by established AI providers such as IBM or Google. Purchasers must be sure to differentiate the IP owned by the target (i.e., those possibly created through using the AI created by the provider) from the IP that will remain with the provider post-acquisition. This is largely dependent upon the contractual relationships the target has with its customers and the AI provider.

Regulatory landscape

The Canadian government has been committed to providing significant monetary support to the AI industry. In the 2017 Federal Budget, the Government committed $125 million to develop the Canadian AI industry through the ‘Pan-Canadian Artificial Intelligence Strategy’.

However, despite this financial contribution, Canada has not created a comprehensive AI policy framework. There are many common legal questions that arise with AI: who owns the copyright of a painting created by an AI? Who is responsible when an AI driven car causes injury? The legislative framework to answer these questions is remains sparse.

The Canadian government will be expected to fill in the regulatory landscape affecting the AI industry in the coming decade as it attempts to navigate these stormy waters. The changing of regulations governing any given industry will cause associated risks to those companies operating in the field. Fortunately, the government’s stated dedication to becoming a world-leader in AI indicates that these changes will mostly be positive for the industry. However, when making investments in AI companies, buyers should be aware of the legislative and regulatory tenor regarding the AI industry as a whole.

The author would like to thank William Chalmers, articling student, for his contribution to this article.

Stay informed on M&A developments and subscribe to our blog today.