In this episode of Techopia Live, OBJ publisher Michael Curran interviews Brian Kells, a corporate and securities lawyer from Perley-Robertson, Hill & McDougall, discussing the role of AI in the tech sector and its legal ramifications.
The following transcript has been edited for clarity and length.
TECHOPIA: Brian, can you share a little bit about yourself and what you do at Perley?
KELLS: I’m a lawyer at Perley-Robertson, practicing a broad range of corporate, and commercial securities laws, all different areas of law that are important to businesses in the Ottawa area. I’ve been practicing for about eight years now at Perley, and we’re the largest independent firm in Ottawa. We have lawyers offering full-service scenarios, including employment law, intellectual property, commercial litigation, and real estate. Essentially, if you’re a business in the Ottawa area, we’re able to assist you.
TECHOPIA: Brian, technology evolution flows in these big waves, these giant trends. There’s a current wave, you might even call it a tsunami, that’s artificial intelligence. Given your experience in corporate law, any general thoughts on where we’re at in terms of AI these days?
KELLS: Technology does go in waves, and this one in particular is pretty exciting. I think AI has the potential to be very transformative. Already, it’s incredibly useful. You’re able to go on ChatGPT or other AI providers, and already it’s providing value to people and you’re able to use it, but we’re just at the start of it. I think ChatGPT 3.5, or whatever came out about a year and a half ago, and we know they’re already working on more advanced models. If you look forward a little bit, what they’re shooting for is artificial general intelligence. If that were to come about, then it’s going to open up a whole new wave of opportunity for businesses.
TECHOPIA: Brian, give us the dos and don’ts of AI?
KELLS: Some of my dos might be a little bit on the cautious side as well, being a lawyer, but I’ll just let you know some of the things I think about when I’m using it and when clients come to us for advice. One of the big things, and we’ve already advised a few clients on this, comes from the perspective of when you’re partnering with an AI provider to help train their model. The way these artificial intelligence models work is that they rely on really large data sets that they turn through to train their model so that then they can work on it. Some of these companies have been approaching businesses that have data themselves to run the program on it and to train on it. Anytime you’re talking about that, if you’re a business that has a lot of data, you have to consider Canadian data privacy laws.
For example, if you’re an existing company with a large volume of personal information, customer information, spending habits, etc., you might want to partner with an AI company, like Amazon, Apple, or OpenAI, to bring their model in-house and to look at all of your data and then see if it can have any insights in that. We just have to keep in mind Canadian data privacy laws. The principal one is the Personal Information Protection and Electronic Documents Act, PIPEDA, which imposes several obligations on businesses that retain personal information and data. You have to make sure that what you’re doing is going to comply with that.
TECHOPIA: Brian, what are some of the don’ts when you think of AI?
KELLS: One of the big don’ts involves copyright of data. As an example, let’s consider the Ottawa Business Journal. It has 20, 30 or so years worth of articles that it has published. That might be really useful data for an artificial intelligence company to run its algorithm on it and to train on it. And as the owner, you would want to make sure that when you’re handing that data over, you have the full rights to do that. What we’re talking about really is the copyright that exists in some of those articles.
If you’re someone in that situation, if you have employees that have been writing things for you or publishing things for you, that’s a clearer sense of ownership. If they work for you, then you generally have the copyright to that and it would depend on the employment contract, but if you’re in the situation where you’re bringing in outside authors, freelancers, that sort of thing, it’s not necessarily clear who has the copyright ownership to that material. If you’re going to be partnering with an AI company, you want to make sure that the data that you’re handing over, let alone the data privacy that we just spoke about, but you want to make sure that you have the intellectual property rights to be doing that.
TECHOPIA: Thank you for bringing that one home, you made it very real for me. It’s no longer a theory. You’re right, do we have the IP and copyright on material we could both share? That’s a great do and that’s a great don’t. The other thing I’ve talked to some of your colleagues about that we could touch on just briefly is, just be cautious. There’s a temptation to let AI draft everything, to draft a sales agreement, an employment agreement, or a contract with a major supplier. AIs can be useful, in drafting some things, maybe more on the marketing or content side, but talk to us just for a minute about the danger of actually getting AI to draft legal agreements.
KELLS: The main takeaway is not to trust it too much, especially in areas where accuracy matters. Like the example you use, a marketing plan, or even content on LinkedIn. When accuracy matters, you have to make sure that you’re reviewing what it does to ensure that’s right. The way these models work is, you’re asking them to make things up. There are examples where they tend to fabricate things where accuracy matters. An example that we talk about in law a lot is, recently this happened down in the United States, there was an attorney who used an AI, it was OpenAI’s model, to generate his pleadings before the court. Then he submitted it to the court, but it turns out that the model fabricated some case law. They were fake cases, they never existed. And he put that before the court and got in a lot of trouble.
TECHOPIA: Thanks, Brian.
