Human Resources Committee on Nov. 1st, 2023

[Michael Coteau, Liberal, Don Valley East, ON]

Thank you very much, Mr. Chair.

I want to thank all of the witnesses who are joining us here today. I think this is an incredible opportunity for this committee to learn from you, and for you to share valuable information to help guide policy-makers going forward when it comes to the labour market.

I'll start with Ms. Lane.

Near the end of your presentation, you actually started to talk about the role of policy-makers in AI in general. Where do you see policy-makers? Could you maybe spend a bit more time on the role you see policy-makers playing in contributing to the growth and adoption of AI and the expansion of AI into our workforce?

Maybe you can just tap into some of the jurisdictions where you see it working well, and perhaps provide some insight on some of the key points that we should always consider when moving forward.

[Marguerita Lane]

I was really pleased, hearing some of the other presenters, that there actually seems to be a lot of consensus in what we're seeing. There seems to be a pretty consistent picture emerging in terms of the benefits but also the dangers of AI.

Let me get to this question of what policy-makers should be doing. I think the first thing is that you want to reap the benefits of AI, and you want to ensure that those benefits are available to everybody and to every company, as well as every worker. You also want to address the risks that AI poses for workers' fundamental rights and their well-being.

I began to describe how existing legislation, of course, is there to deal with some of these risks in terms of legislation for discrimination, data protection, workers' rights to organize, and occupational safety and health. I think that's already a very important starting point. We don't throw those things away just in the face of a new technology. Of course, there might be a need then for some AI-specific legislation on top of that, and for soft law to adapt as things move forward.

You asked a question about what is working well. I think the firms that are doing well working with AI, getting the benefits out of it and allowing their workers to benefit from it as well, tend to be those that are training their workers to use AI and consulting their workers about AI as well—this was mentioned by a previous speaker—and where workers can participate in the process. I think those are two key factors.

I also think public policy for training and education are very important to address changing skills needs, and then, of course, there's a need to ensure that collective bargaining and social dialogue can play a role as well in supporting workers and businesses through the AI transition.

I think it's also very important, as I said, for everybody to have the opportunity to benefit from AI. This goes back to education and training, but equally it's important to ensure that, for example, workers in SMEs also have the ability to adopt AI, and that they aren't excluded from the process just because of the data requirements and the heavy IP requirements as well.

For policy-makers, there's a requirement to keep track, to keep on top of all of this, so as I mentioned, the OECD's efforts are really focused. For instance—

[Michael Coteau, Liberal, Don Valley East, ON]

I'm going to jump in. I think I have about one minute left, and I really want to ask some of the other witnesses some questions. Everything you said is very valuable information, and I wish I had more time, but I'm going to jump to another witness, and I do apologize for interrupting.

Mr. Carbonneau, you talked about IP. I know that the CCI has taken a position in the past that a lot of the intellectual property that's coming out of our universities is removed and brought to other jurisdictions because there's capital there. There may be more flexibility in those jurisdictions to scale up quickly.

One of the criticisms I've heard around why AI doesn't have the potential to grow in jurisdictions like Canada as much as it would in the United States or China is that it is because of the datasets that are being used to actually provide the input into AI for machine learning. Is that necessarily true, from your perspective?

[Laurent Carbonneau]

That's a good question, and I think I would have to consult with some members to give you a very hard answer. What I would tell you is that this hasn't come up.

I would say, perhaps on the processing side, that access to compute is an issue such that perhaps we might hit a bottleneck there in ways that other countries might not. I believe the government is perhaps looking at some options around that, which is, I think, all to the good.

I think we've had excellent efforts, like Canary, in the past, in terms of having some public compute capacity. I think that's a good idea and definitely something we should try to avoid that bottleneck around.

[Michael Coteau, Liberal, Don Valley East, ON]

With respect to the 150 tech companies that are part of your consortium, there have been hundreds of millions of dollars invested by Canadian citizens, by government, into supporting research into AI. Do many of these companies benefit from that type of research investment?

[Laurent Carbonneau]

They do, absolutely. I think where we fall apart sometimes is in just making sure that Canadians see more benefit from the public investment we put in. We do get some—that's undeniable—but we want to make sure that the proportion of public value for public dollar to Canadians is as high as it can be.

[Michael Coteau, Liberal, Don Valley East, ON]

Thank you so much.

[The Chair, Liberal, Bobby Morrissey]

Ms. Chabot, you have the floor for six minutes.

[Louise Chabot, Bloc, Thérèse-De Blainville, QC]

Thank you, Mr. Chair.

I want to thank all the witnesses.

Ms. Lane, I want to make the most of your presence here as a representative of the OECD. Thank you for joining us. I know it's late.

I have two questions for you.

I asked people with the union movement in Quebec who represent workers in these sectors specifically to come testify for this study. AI technology may not have a tangible impact on labour yet. It's theoretical. We're trying to foresee what might happen, but nothing is clear yet.

Anyway, in Quebec, the questions people are asking about AI have much to do with ethics, especially in higher education. Many studies have been done on the impact of AI on work ethics and production ethics. People in some sectors are even saying we should put the brakes on. I think it will continue, just like automation.

We've talked about laws, but what should we be doing right now to get a glimpse of what's going to happen, kind of like what happened with automation? How can we get ahead of things given that the impacts are not very tangible yet?

[Marguerita Lane]

You talked about how the impact of AI at the moment is not so concrete; it's more theoretical. I think we are at an important point between the theoretical and the realized. In our survey, for example, when we asked workers about the impact on job quality and when we asked employers about the impact on job quantity, we were really asking about what has happened already. We found that there were already changes happening in the manufacturing industry and the financial industries. Yes, absolutely, much of the impact is theoretical, but I think it's important to note that there are changes happening already.

I think that there's an important point here about ethics and that there are important ethical questions related to AI. Within the workplace, we already have tools for dealing with some of the issues that AI raises. For instance, we can talk about the ethical risk in terms of the biases that AI can introduce, but, within the workplace, we have anti-discrimination legislation, so I think we can frame things in terms of ethics. This makes them sound a little more theoretical, but I think we do have real tools already at our disposal to deal with some of the more theoretical or more ethical aspects of AI.

What's an action that we can take now? I think we can start thinking about how we want to use AI in our society and in our workplace, and what we think are acceptable and good uses of AI. That can be a starting point for legislation around AI.

Just because something can be done, just because something is feasible within the technology, it doesn't mean that we have to follow through with it. Whether it's as a society or unions or businesses deciding this, perhaps all together we can already start to talk about what kind of society we would like to have and what AI's role in that society should be.

Thank you.

[Louise Chabot, Bloc, Thérèse-De Blainville, QC]

You talked about the European Union, which I believe passed a law. Are there other similar examples?

I'll leave it there for labour, collective agreements and unions.

What is a government trying to govern by enacting a law on artificial intelligence? What is the legislator's motivation?

[Marguerita Lane]

In the EU's AI Act, they have chosen to take the approach of essentially looking at high-risk and low-risk AI. The idea is that high-risk AI is the type of AI that will have to go through a specific process, a more burdensome authorization process, whereas low-risk AI can essentially sail through.

High-risk AI could be AI, for example, that's being used in situations where there is the risk of harm to a person—for example, their livelihood, their freedom or their health. That's essentially one approach that's being taken, and we will see how it works in practice.

Another interesting example is Germany. This has nothing to do with AI at all. This has been in place for a while. They essentially mandate worker consultation whenever any kind of technology is being introduced into the workplace and impacts workers' working conditions. I think it will also be interesting to see how that existing system is able to respond to the introduction of AI. It's not a piece of AI-specific policy; it has been in place for a while.

Those are two examples of how people across the world are dealing with this challenge.

[The Chair, Liberal, Bobby Morrissey]

Thank you, Ms. Chabot.

[Louise Chabot, Bloc, Thérèse-De Blainville, QC]

Thank you very much.

[The Chair, Liberal, Bobby Morrissey]

Madam Zarrillo, you have six minutes.

[Bonita Zarrillo, NDP, Port Moody—Coquitlam, BC]

Thank you, Mr. Chair.

I think the majority of my time on this round will be with Mr. Carbonneau, please.

Thank you to all the witnesses.

I want to pick up on something in the testimony of Mr. Carbonneau on net incomes being too low in Canada. This is on my mind as it relates to AI. How do we protect the income of workers when their cognitive value, which is their ideas or their thoughts, is already captured, copied, scaled and potentially monetized?

How do we protect their cognitive IP? We always talk about intellectual property, but how do we protect their cognitive property when it maybe only needs to be a thought that has been shared once, twice or three times and it can be scaled and used?

[Laurent Carbonneau]

That's definitely a big question, and I don't think I have a complete answer. I'm not sure that anyone does.

What I would say is that, in a broad sense, I think countries that do well in AI are going to be the ones that are able to develop acceptance for AI adoption and use in societies, and I think that we will have to answer those questions in some format probably sooner rather than later.

We do have a bill before Parliament right now, Bill C-27, that is implementing a legislative framework to develop a regulatory framework around AI. I think there's a lot of scope there, as that comes into force and the regulations are developed, to be quite sensitive to what the future of those kinds of issues looks like.

I will applaud some of CCI's other work here. We released a road map on responsible AI leadership in, I think, early September—time has blurred this fall, as I'm sure it has for many of you—that really gets into some of these issues around public trust.

I think one thing Parliament should strongly consider moving forward is creating a parliamentary science and technology officer who would play an analogous function to what the Parliamentary Budget Officer does and very similar to what the sadly now-defunct Office of Technology Assessment used to do in the U.S. Congress. It would give you as parliamentarians and the public timely, actionable information on emerging technology and science issues that would help inform a lot of these debates and give us all a level ground to understand a lot of these emerging technology issues.

I think that's the kind of social infrastructure, if you will, or parliamentary infrastructure that could play a very helpful function in addressing those kinds of issues and give us, I think, a better basis to do so.

[Bonita Zarrillo, NDP, Port Moody—Coquitlam, BC]

Thank you for that.

I thank you for your submission, too, the written submission, because I know that trust was one of the things that went into your written submission.

I think that one of the challenges with that is whether the average worker, the average family in Canada, has access to conversations around that. Do they have access to the decision-making process around that? I think that's one of the things that plays on the trust. What is difficult, is that, yes, some other body might be created, but how do they get access points into that?

It gets me thinking about how that interacts with the conversations that are happening right now around basic income, because it's almost as if workers will have less power, especially if a lot of the value of this AI technology is in servers that aren't in Canada, or the data isn't in Canada.

Have there been discussions around how basic income intersects with artificial intelligence as it relates to incomes for people?

[Laurent Carbonneau]

From our perspective, I'm aware that there are some. This has been a topic of discussion at a high level, but speaking for our members, I can't say that this has come up a whole ton for us, because I think this is a fairly remote prospect in terms of being a reality.

[Bonita Zarrillo, NDP, Port Moody—Coquitlam, BC]

All right. Thank you.

It does worry me when I hear that we're consulting with manufacturing and finance when we have a lot of women in the care economy. A lot of women workers who do care aren't, it sounds like, getting equal exposure in this conversation.

I'm going to move over to Mr. Roberts.

Thank you so much for your testimony.

One thing that is on my mind—as this committee also looks at persons with disabilities—is that there's potential opportunity here for more equity in the workforce as it relates to persons with disabilities.

I just wonder if you could expand a bit on how this could assist workers. I think also about marginalized workers in the care economy, who have been undervalued, under-resourced and under-protected in the economy.

[Chris Roberts]

I had hoped to address this in my introductory remarks, but I didn't have time.

I think you're absolutely right that for workers with physical challenges, for instance, AI applications have enormous potential to increase their participation.

I think the surrounding investments in workers with disabilities are absolutely determinative as well. There has to be the openness to accommodating those workers. The arrangements and conditions that unions and disability rights advocates call for now also have to be included in the mix.

There's no simple technological solution to the barriers and challenges that workers with disabilities face, but there's no question that there's potential in the technology. It's all about how the technology is adapted, deployed and implemented in workplaces.

[Bonita Zarrillo, NDP, Port Moody—Coquitlam, BC]

Thank you so much. I just want to ask about conversations about incomes.

Right now, incomes are too low in Canada. We know that. We know that even income supports aren't keeping up with the cost of living in this country.

Does the Canadian Labour Congress see artificial intelligence being able to enhance incomes or support more income protection programs? How do you see this affecting the incomes of Canadians?

[The Chair, Liberal, Bobby Morrissey]

Give a short answer.

[Chris Roberts]

The technology is not determinative. It's the conditions and the social structures that matter.

That's why I don't agree that productivity is determinative in the short run. In the long run, it may shape and determine prosperity, but if productivity advances from AI flow entirely to the owners of these applications and the firms that deploy them, then it's not going to improve incomes for the lowest income-earners.

[The Chair, Liberal, Bobby Morrissey]

Thank you, Madam Zarrillo.

Before we go to the second round, I just want to advise Ms. Lane.

It's your option to stay. You can see that a lot of panellists want to speak to you. When you choose to leave, simply wave your hand and log out. It's your call, Ms. Lane.

We'll go to Mrs. Gray for five minutes.