The overarching supposition of Artificial Intelligence (AI) is its promise of a better and more efficient life. What does that mean? What health benefits could AI offer? How could AI be used to improve the health outcomes of Canadians?
Professor Steven Hoffman, the Scientific Director of the Canadian Institutes of Health Research’s (CIHR’s) Institute of Population and Public Health based at York University, sat down with Brainstorm to discuss this promising idea. Hoffman, cross-appointed between the School of Health Policy and Management and Osgoode Hall Law School, reflects on the unique role that York could play as Canada takes the lead in AI.
Q: How could AI be used to improve the health outcomes of Canadians?
A: AI is all about taking advantage of the fast pace that computers can process information to draw insights. It can help detect threats before traditional mechanisms and help us understand what kind of interventions make us healthier. Since computers can access and analyze data on a scale that boggles the human mind and AI can identify patterns from data drawn from many different sources simultaneously, it means that, potentially, our insights could draw from a much wider set of information.
For health, this is particularly important because so much of our health depends on complex interactions of many factors. Computers have the potential to incorporate much more information to yield insights that could be helpful to us.
Q: How could AI assist in monitoring and tracking the spread of the flu virus?
A: We have some traditional methods to track the flu. But how much better could it be if we could draw upon a much wider range of data to see the signal of a flu outbreak starting as early as possible? That wide range is not just about hospital or doctor visits, but also things like what we type into Google or what we post on Twitter. This full range of data would be advantageous.
“Canada is in a leadership position in AI. We need made-in-Canada AI applications for public health.” – Steven Hoffman
Q: How could AI be used to predict the next epidemic?
A: One of the most important things in an outbreak is the ability to respond quickly and appropriately. Right now, one of the challenges is to identify an outbreak as soon as possible. If you look at previous outbreaks, there’s quite a gap between when unusual patterns of illness start appearing versus when the world detects it, and even further delays before the world is able to respond to it in any meaningful, coherent way.
So, we need to develop innovations to minimize those time periods. Every day that an outbreak is developing unchecked increases the potential for a true pandemic, and raises the human and economic toll of the outbreak.
Q: How could AI be used to address the social determinants of health (SDOH), such as poverty, housing and food insecurity?
A: I’m particularly excited about AI in the context of the SDOH because these determinants encapsulate different dimensions that aren’t always comparable with one another. For example, income is expressed in dollars, while food insecurity is expressed in calories or nutrients.
In traditional models, there has been a tendency to focus on individual SDOHs, which then can lead to a siloing of the policy response. But the SDOH are highly connected; they all depend on each other.
If we develop models that consider multiple different types of data, as AI facilitates, then this could allow us to understand the SDOH in a much more profound way, including the ways we could intervene.
“York is extremely well positioned to unpack these problems and offer solutions in a way that will allow for social justice and equity to be achieved. York could take a thought leadership role on making sure that AI works for everyone.” – Steven Hoffman
Q: Is Canada in a leadership position?
A: Canada is in a leadership position in AI. There’s also some great research efforts here in the healthcare sector. Where we’re lacking is at the interface between AI and public health. That greatly worries me. That’s why CIHR’s Institute of Population and Public Health is making the use of AI methods for public health challenges a top funding priority.
Our concern is that there are other actors, like tobacco companies or junk food and soft drinks manufacturers, that might already be using AI methods to identify future customers. We’re seeing some companies coming to the battle with the big guns [big data], yet public health is bringing a butter knife in its response. There’s no competition.
We need to make sure that Canada is equipped with the research expertise to use AI methods to address public health challenges. It’s really important. We need made-in-Canada AI applications for public health that are reflective of the unique Canadian context.
Q: What is York University’s contribution?
A: One of York’s great strengths is its emphasis on social justice and equity. When you think of AI, there are lots of social justice and equity problems. You see, the leading form of AI is machine learning, which uses massive amounts of data to identify patterns. The challenge, here, is that a lot of our data involves middle- and upper-class men to the exclusion of women and lower socioeconomic status groups. As a result, these AI algorithms bake in the inequities. We could end up with a system that only reflects, and serves the needs of, a small proportion of society.
York is extremely well positioned to unpack these problems and offer solutions to them in a way that will allow for social justice and equity to be achieved. York could take a thought leadership role on making sure that AI works for everyone.
To learn more about Hoffman, visit his faculty page or the Globe & Mail opinion piece (January 22, 2018), cowritten by Hoffman and CIFAR President, Alan Bernstein. Also noteworthy: Osgoode held an AI conference in February 2018, Bracing for Impact: The AI Challenge.
By Megan Mueller, manager, research communications, Office of the Vice-President Research & Innovation, York University, email@example.com