How transformative does AI stand to be for the business sector, and for Canada’s economy in particular?
I believe that AI is going to be hugely disruptive. It is hard to say exactly on what horizon since this depends on many technological and research factors. However, even if research did not make any additional progress, a huge amount of benefits will emerge from the deployment of the current science across all sectors. Such progress would simply follow from larger databases to train those machines, as well as from the emergence of bigger and faster machines, as the leading chip manufacturers are building specialized neural net chips and deep learning chips these days. So these developments and all the engineering that is happening in companies and in academia will lead to many applications across all sectors, and AI will impact the Canadian economy across all sectors.
In order to determine which sectors will be impacted more than others, one key factor is the availability of data, which is going to be a necessary condition to the application of AI techniques. As such, the decisions companies are taking now about how they collect data and what data they collect will have an impact on their AI applications in three, four, five years from now. Sometimes the issue is not just how we are going to integrate AI into companies or industries, but how we lay the groundwork so that we will be in a good position to be winners a couple of years from now. The industrial sector, which already has some automation and data collection, is a very good example of an industry ripe for AI disruption. The health sector is another industry with lots of potential since there is a lot of data there; we just have social challenges related to how that data should be shared properly in a way that is safe for patients.
One of the most interesting Canadian industries that will be revolutionized by AI is the tech sector itself. Rapidly growing AI Companies like ElementAI, for example, are contributing to Canada’s tech industry. So our tech industry is going to grow and is going to be fuelled by all of the AI advances being created here. This is something that many countries are investing in right now and Canada must really be a leader in those investments in order to take advantage of the intellectual assets that we have. Canada is really ahead of most countries in this respect.
What can Canada do to build on the momentum that Montreal, Toronto and other Canadian cities have achieved in terms of positioning themselves as global AI hubs, and what must we do to remain internationally competitive?
Canada’s AI hubs emerged due to their large groups of researchers with prominence and visibility in the scientific community. This was achieved mostly in Montreal, Toronto and Edmonton, and in Vancouver to some extent. There are also other Canadian cities that have pretty good machine learning groups, like Quebec City and Waterloo. This talent pool is what has attracted big tech industry players to Canada to open up research labs in those cities where a critical mass of PhD students, graduates and now engineers are found. Our talent pool is a rare and sought after commodity, and one of Canada’s advantages.
In addition to our people, one of the strengths we have is our community spirit. In Montreal, for example, all the AI companies worked together to represent Montreal at the Neural Information Processing Systems (NIPS) Conference, the major machine learning conference. If you attend that event you’ll see many big companies selling their brand. In our case, Montreal companies were selling Montreal’s AI brand; all of those people and all of those companies were working together. There is a spirit of collaboration versus competition, which is stronger here than in other places. Another aspect that differentiates our companies is that they generally reflect the values of Canadians, and they care about the social impact of AI. In that respect Canadian governments are not just investing in the scientists who will build the next generation of AI and start-ups, but they are also pushing us to connect our computer scientists with people in the social sciences and philosophy who care about ethics and law, and how this technology is going to change society. Most Canadian AI researchers, like myself, do care about those questions. They want to ensure that their work, and the products that are going to be built off it, are going to be used for the good of humanity. And this forms part of the Canadian AI brand, which actually helps us recruit people from the US to Canada, for example, especially in the current political context.
In order for Canada to remain globally competitive, part of the answer is to think about Silicon Valley. Silicon Valley is a small place. Not that many people live there but look at the impact it has on the world. So it is not just the number of people that is important. It’s the existence of an ecosystem that feeds on itself in a positive way and that can have a worldwide impact.
I believe Canada can replicate this. Firstly, we need to be able to retain our strong scientific talent because there is a powerful pull from other countries to recruit those people – to drain those brains. Secondly, we also need to train a much wider circle of engineers and people who would be applying that science. Thirdly, we must develop AI start-ups across all sectors. That is something that is not necessarily what Canadians have a strong track record in, so we need to really get our act together and make the right decisions.
We must make sure we keep our best companies here before they get bought by some large foreign tech giant. There are all kinds of issues we must address in order to do so, which are not only specific to scaling up AI start-ups. We must also look at how we marry our academics with business people and industry. This is always a challenge because these groups have different cultures and different horizons. But we must try; we have to invest and we can’t wait. We have to step on the gas and we have to leverage our specific Canadian strengths.
What impact will the recently announced SCALE.AI supercluster have on the industry’s development? Is the government doing enough to support Canada’s AI industry?
The superclusters will bring all the industry’s actors together around the same table. These include the universities, the academic research centers like MILA here in Montreal, the Vector Institute for Artificial Intelligence in Toronto andthe Alberta Machine Intelligence Institute (Amii) in Edmonton. But we cannot do it all through our research centers; most of the change is going to happen in industry through existing companies and new ones. We need all these actors to talk to each other quickly and understand what is going on in the Canadian AI ecosystem.
There are two things that I repeatedly get as a feedback from companies seeking to leverage AI’s potential for their business. The first is about talent: companies need more people that have enough training to help them. Secondly, they need knowledge, they need help, they need guidance. The markets are eventually going to fill those gaps but we do not want the markets to do that on their own as they will take too much time while other countries are accelerating ahead and taking the lead. That is why the government support for the SCALE.AI supercluster is so important to accelerate the Canadian AI industry’s development so that we can become leaders economically in AI and not just scientifically.
The government has now clearly showed that AI is an important part of its strategy and thedecisions that have been taken up to now are great. Also, we are now entering the scale-up phase of the industry’s development, so I think the clusters are great and we now have five clusters that are quite sector oriented. However, what I think we need eventually is a Pan-Canadian AI supercluster that really focuses on the AI technology and its application to all sectors of the economy, not just one sector at a time.
You participated in the creation of the Montreal Declaration for a Responsible Development of Artificial Intelligence. Why is this important? Are governments and companies doing enough to ensure that AI’s development is framed in the appropriate way?
The Declaration is very interesting. It is a set of seven ethical principles on how AI should be used and which are the values that should guide its development. These kinds of principles could be the backbone of legislation or other government action to make sure that what we do with AI is going to be aligned with our values and respectful of our humanity. What is also interesting is the way it was produced. It wasn’t just a bunch of academics talking to each other and coming up with a set of principles, but it involves everyone. Everyone is invited to comment on those principles and to propose changes to them via the website and through events. For example, I was recently at an event here in Montreal where different people were proposing changes, because there may be some aspects that we did not think about. So the Declaration was produced as a kind of a collective construction, or co-construction.
I think this is an important feature because AI is not only going to impact science and the economy, it is going to impact ordinary people. It is going to force politicians to take decisions about how to deal with the changes AI will spur. So it is important that ordinary citizens also understand the issues; not necessarily at a very complicated, abstract level but in the ways that the technology is going to impact them. This will occur in areas such as healthcare, private data, military applications or security applications. There are all kinds of non-obvious choices that need to be made and I think ordinary citizens need to understand those issues.
“Most Canadian AI researchers […] want to ensure that their work, and the products that are going to be built off it, are going to be used for the good of humanity. And this forms part of the Canadian AI brand.”
There could be misuses of AI; people could design AI for purposes that are good for them but might hurt other people. This is part of the ethical aspect of AI’s development. Governments will have to agree with each other to come up with treaties to minimize the negative impact that this technology can have, such as applications related to warfare and security, for example. Also, we need governments to update our social safety net because the transitions AI will bring on are going to happen very quickly. We have a welfare system in place but I do not think it is adequate for the changes that are coming. I am not an expert on these issues but I do think we need to sit down and think about them to see how our social systems may need to be reorganized. This is one reason why so many people in the tech industry want to see some kind of universal basic income. They do not want to feel responsible for having contributed through their research to so many people losing their jobs and being in misery.
How will AI disrupt the future of capital, labour, work and education?
I think in the long run our relationship to capital, labour, work, and what work is, are going to change drastically because of all of AI’s impacts on society. It is very obvious now that the value of labour is going to decrease and the value of capital is going to increase because it enables the purchase of those machines that are going to do the work that workers used to be doing.
There is a risk that many jobs will be lost. It is hard to predict actual numbers but the impact will be significant. In the shorter term, the impact of AI on the workforce will be lower. But in the longer-term we might see a profound transformation. I think that over a longer period of time we will have the opportunity to build a society where everybody can contribute.
In particular, I think there should never be a shortage of people taking care of other people. We can have more teachers, more nurses, more people taking care of babies. All of these outcomes are positive, and we do not want robots for those jobs. We will want our elderly parents to have a human around them when they are in the hospital, or a baby to have a human around her when she is being taken care of. People can be assisted by technology to make such work more efficient, safer and better in many ways. But humans are crucial for the many jobs where there is an emotional relationship. Eventually, we will have many more people doing human-to-human jobs. And the jobs in manufacturing and big industry will probably go.
These shifts are already happening. So we need to train and educate people who will have more general knowledge and who can therefore adapt to changes faster, because it is hard to predict which jobs are going to disappear first. We want to make sure that people are more adaptable, that they can learn and reason by themselves. These kinds of skills, like human relations and leadership, are not widely taught to our high school students right now but we could do more of that. Of course, if one studies in an area where you are going to be among the people building those robots, then yes, that is good, but not everyone can be doing that.
Our relationship to work is going to change because it is not simply negative that industrial jobs are going to be replaced, it is also positive in some sense. What I would like to see in the future – and I think it is very possible – is a scenario where our work is not something we do because we need to earn a salary to survive. For most people right now, working is a kind of slavery. The vast majority of people do something that they would not do if they were not paid to do it. There are many people in society who fall into this category, but there could be many more who do something that makes them grow as people and feel like they are useful and creative. I like to think of it like my own work where I feel so lucky that there are people paying me to do something I would do anyway if I could – not because I need the salary, but because I really want to do it for its own sake. I enjoy it. I feel that I contribute and that it makes me a better person. So this already exists but it can grow and make more people happy and feel good about their role in society.
This article originally appeared in The Future Economy