Format Légal | AI success in an evolving legal context: What you need to know

INTERVIEWER: Hello, and welcome to a new episode of Formal Legal, the legal news podcast from Gowling WLG. Be sure to follow us so you don't miss an episode, and visit our website at gowlingwlg.com for all our resources and the latest legal news.

We live in a world where digital technologies are evolving at high speed, and ready or not, many businesses are being called upon to make the leap into this new era. When we talk about new technologies, we're talking about artificial intelligence, which is making its mark on our daily lives, whether through ChatGPT or any other application.

AI is transforming many sectors, and while the potential benefits are immense, the legal implications can be overlooked, misunderstood, or treated as isolated issues. This episode aims to set the record straight and provide you with concrete, practical information on the crucial legal questions that businesses need to ask themselves in this age of AI.

We will look at how a holistic legal approach can help companies navigate the regulatory framework, protect their intellectual property, address governance issues, manage risk, and ultimately succeed over the long term. So are you ready to decipher the legal intricacies of AI success?

Today we're talking to Denis Keseris, partner and patent agent in our Intellectual Property Group in Montreal, and Naim Antaki, partner and of co-head of the National Artificial Intelligence group at Gowling WLG. Both are also members of the technology sector group in Montreal. Hello, Dennis. Hello, Naim.

DENIS KESERIS: Hi, Delphine.

NAIM ANTAKI: Hello, Delphine.

INTERVIEWER: So, to begin with, I would like to know and if you can tell us about the current state of AI, in other words, where are we in the latest developments in AI, and I know it's moving fast, so it's going to be as of today.

NAIM ANTAKI: It is moving very fast. AI is on, everybody's minds. It's the subject of every conference and it's also the subject of a lot of boardroom discussions. And I think it's for good reason, because this is truly a technology that has the potential and that has already started in some cases to make a real impact on how people do business and how well they can do it.

The key thing here is that it is constantly evolving from a technological standpoint. And we're still figuring out certain things. At the same time, law is trying to catch up and law can only catch up because the last thing that you want is for a law to be implemented on a technology that is not yet mature enough.

And so this is where you always have this difficult balance between people who want to have pure clarity on what they can and cannot do from a legal standpoint. And the legislator who is saying, well, I don't want to break things or stop an industry by regulating too fast. And so it's quite interesting for us, very exciting for a multidisciplinary group to navigate in this with our clients.

DENIS KESERIS: Yeah, I think, and from a technology perspective, as a patent agent and chartered patent attorney, I've been looking at technology now in my career for 20 some odd years. And I've seen a lot of new technologies come out, from nanotechnology, to graphene, to virtual reality, and augmented reality. I think AI is becoming apparent that AI is actually a revolution in terms of technology, and I think you can almost measure that by looking at the number of sectors that are going to be transformed or are being transformed by artificial intelligence and the depth of that impact.

In some industries, I think we're going to see some very, very large changes.

INTERVIEWER: True. Thank you. It's true that artificial intelligence raises many, many questions and issues. Naim, could you give us an overview of them and above all, tell us why these issues should not be treated as isolated questions. In other words, how should companies approach risk management in the field of AI.

NAIM ANTAKI: I think the important thing is to make sure that everyone is aware of the technology itself and the fact that there are different technologies when you talk about AI and that it really is an umbrella term. That's number one, because you can only properly evaluate the legal impact of something when you understand the facts. Just like you would in a strategic setting, et cetera, et cetera.

Now, from a legal setting, a lot of people have been focusing on AI specific legislation, legislation that has AI in its title. So the EU AI Act is one of them, or the Draft Artificial Intelligence and Data Act at the federal level, which is still being discussed in committee and is not yet law. That's OK, but that's a unfortunately a limited regulatory view on what the true broad impact of AI truly is.

And so you also need to think about laws of general application, where you may start seeing either small changes to the law itself or guidance coming from regulators, or decisions coming from judges that apply questions relating to privacy, questions relating to liability, questions relating to employment. And so it's very important when you think about this to also think about the fact that a gray area in a particular area of law could be solved, not within that area of law, but perhaps through a different area of law.

And as a business lawyer, I always try to think about how contracts can be the perfect tool to try to allocate risk between companies and frankly to be clearer about who should do what, and if something goes wrong, who is truly liable. Because one of the very first questions that AI companies large and small ask us is, well, what happens if something goes wrong? Am I liable? And is it really my fault or not?

And I think on this, what's really important is to make it clear in the contracts and to really negotiate the contracts, but also to be transparent as to not just the opportunities but also the limitations of the specific technology that you are using with your customers, with your employees, with other people, so that people know what they're dealing with. And if they know what they're dealing with, they won't feel slighted if a limitation that you knew existed and that you mentioned could occur actually happens.

And so the goal is to advance. The goal is to grow the business, not to deal with litigation at every step of the way.

INTERVIEWER: And so AI raises a number of intellectual property issues. Denis, can you explain why.

DENIS KESERIS: Yeah, well, I think there's a lot of change that has happened in the last few years on the legal side of AI and that is the case in intellectual property as well. There are some issues that seem to be pretty much settled, patentability, for example, while there has been some activity in the courts around the world in relation to the patentability of artificial intelligence. Most jurisdictions have case law that is fit for purpose in assessing whether or not an AI implemented invention is patentable or not.

The other one recently there's been a lot of activity around the world on is whether or not AI or whether or not an AI can be considered as an inventor. I think also that's pretty much settled in most large jurisdictions around the world, the answer is no. And then there are other aspects of intellectual property law that remain maybe less clear, one of which is copyright, and whether or not the use of copyrighted material or materials in which copyright subsists can be fairly used to train machine learning models.

And that is still before the courts in most jurisdictions, and we're going to have to wait and see how that plays out around in courts around the world.

INTERVIEWER: Sure. And what advice would you give to have the best approach or strategy in terms of IP currently related to AI?

DENIS KESERIS: Yeah, I think we need to keep in mind that the situation on some aspects is evolving. We need to understand that some of the calculus, the fundamental calculus in terms of assessing whether which aspects of a company's technology might be patentable and which aspects might be kept as trade secrets, for example, that calculus is very different for artificial intelligence, we need to bear that in mind. But ultimately, we need to also not throw the baby out with the bathwater and make sure that we keep the focus of any portfolio development squarely on commercial aspects.

So patents, trademarks, all of these intellectual property rights are commercial tools, and we need to make sure that in developing portfolios of those assets, we're squarely focused on the commercial aspects of how this technology is going to enter into the market, how it's going to be used and by whom.

A good example of that is-- we have some clients that use publicly available data and off the shelf models. So they train off the shelf models using publicly available data. So some of their core technology might not be patentable, and so we're in a situation where your IP strategy might not be focused so much on patenting core fundamental technology to the business, but more on, for example, OK, how do we protect the user experience? How do we protect our brand? So some of the-- again, some of these strategies are going to be different for artificial intelligence.

INTERVIEWER: Sure, it will require flexibility. Naim, you mentioned that at the forefront that we need to ask the right questions about AI governance. But what kind of questions should we be asking now? And would you have any advice on the direction to take?

NAIM ANTAKI: It's an important question for every organization to ask, whether you decide that you don't want to use AI at all or if you decide that it may be a good idea. Why, because AI is all around you, both perhaps people that you collaborate with and that you interact with who may be using AI. And so you need to understand what's the impact on your business.

Number two is could have bad actors, hackers, who are using the tools that are available now with AI to clone voice, to clone a video, to really impersonate people, in a way that was not easily attainable before. And so, to give you one example, authenticating by voice is no longer as secure as it used to be a few months ago.

So I think the idea here is to always try to think about what's an alternative non-technological way to confirm the identity of a person. And I think more generally when you're thinking about these issues, both defense and offense with respect to AI, it's important to keep in mind that you need to think about talent. You need to think about compute access to computing power, which is not as easy as people think, and was a key focus of the federal government's budget announcement last year, 2 billion out of the $2.4 billion.

And also, of course, data. And as Danis was mentioning before, it's not because data is available that you necessarily have the right to use it. And if you're particularly if you're thinking about personal information, for example, you can only use it if you have consent for a particular purpose and for a reasonable amount of time. So it's not because you sourced a data set in a completely different context that you have within your organization that automatically you can use it for whatever it is.

And so you need to think about the interaction between these different issues. And I think at the end of the day what's most important is constant communication amongst business, tech, and legal, in order to really understand where we are at and what are the ways in which we can deal with risk in a way that makes the organization comfortable.

And when you do this, you really have to-- I think, align with the values of the corporation. You may have certain values that will break a few things, our clients know it, but we want to move as fast as we can. Whereas for others they may want to move a bit more slowly or in a more precise pilot projects, because customer trust is at another level, if I can put it this way.

And so in this whole context, I think it's important to think about AI, not just as a point in time analysis, but really as the initial analysis being just that, the beginning. And as you advance and you monitor the response and going into the market of your solution, how are people reacting are we seeing certain issues that we didn't foresee before. Are we monitoring for that? Are we testing for it? And are we allowing people to make it better?

Another aspect I think that's really important and to always keep in mind, is a concept that is more and more at the forefront in current legislation, which is the concept of transparency. And transparency sounds really great and really easy until you try to think about, how am I going to explain AI in a clear manner to someone who doesn't know what it means? And how my solution can have an impact on her, on him, on them.

And there's always this interaction when Denis and I work on files together. On one hand, you do want to be transparent. On the other hand, you don't want to disclose trade secrets or really key sensitive information of the corporation that makes it that you spent a lot of time and money to develop a solution that the person down the street can just copy.

And so there are a lot of nuances to think about, and I think it's by continuing to really take this holistic approach to AI governance that you're able to advance in a way where you feel comfortable that you may not have it right. But even if you get it wrong now, you can self-correct and adjust your course and continue advancing forward for the long term.

INTERVIEWER: Sure. Thank you, Naim and Denis. Did you have anything to add in terms of takeaway from this rather short conversation on a very broad and important topic?

DENIS KESERIS: Yeah, I think, in seeking legal advice in terms of artificial intelligence, I think we should take a cue from the technology itself. The development of this technology is complex and it's multidisciplinary. And I think having an integrated legal approach from the business side, the tax side, the data and privacy side, the IP side, having an integrated approach to perceiving or seeking that legal advice, I think is a really important aspect.

NAIM ANTAKI: I agree. I think you need to think about privacy. You need to think about your employees, your customers, but you also need to think about sales. You need to think about marketing. And so there's this balance that is really important. And AI sounds very almost theoretical ether. You don't know what you're talking about.

And I think that can be a blocker to some kind of something that slows you down when you're trying to think, well, how am I going to come up with an AI policy like we have, for example, at Gowling already and we help other clients with. If you go with one or two use cases, I want to use biometrics for a particular purpose in a specific sector, in a specific jurisdiction, then things become clearer in your mind.

And then it's easier for the person who helped in the development of the solution to say, wait, there could be a problem here that the lawyer couldn't think about because he or she doesn't have that expertise. Or someone in strategy that says, wait, we could have a problem with capturing the value of this solution because we see that there are certain issues with respect to talent and compute and data that perhaps we hadn't considered in the R&D phase because we didn't have to do this at scale yet.

So they're really, really exciting issues to think about, and the best way to do it is together.

INTERVIEWER: Great, and I think that will conclude this discussion today. So thank you very much, Denis and Naim, for these valuable tips and practical tools for understanding how the legal world is evolving and positioning itself with artificial intelligence. If you have any questions related to AI in this evolving legal context, please do not hesitate to contact Naim/Denis or a member of our artificial intelligence and technology group in Montreal or in any one of our Canadian and international offices.

Thank you again for listening to this episode of Formal Legal. Do not forget to follow us so you don't miss the next episode. And to find out more, visit our website gowlingwlg.com to consult all the resources available on the subject.