Greg Mori is research director at Borealis AI’s Vancouver research centre and director of Computing Science at Simon Fraser University.
Depending which forecast you subscribe to, the global market value for artificial intelligence is projected to reach anywhere between US$36.8-billion and US$1.2-trillion by 2025. This for a series of technologies that only reached mainstream consciousness a few years ago.
At the vanguard of this push are the tech giants that have seen explosive growth thanks to the most lucrative applications of this technology: entertainment and advertising. The result of these investments has been phenomenal changes in our daily consumption of media.
However, in the excitement over what these technologies can do for immediate wealth generation, there’s a risk of overlooking large swaths of the economy that need more of the positive effects that serious AI research can bring.
The world will make much better use of AI if machine-learning capabilities are strengthened across different sectors, particularly ones that can already begin to solve difficult problems that are holding back quality of life and harming the planet.
Serious AI touches areas as vital as transportation, environmental science and health care. But at the moment, much research is following the analytical gold rush spurred on by search, social and recommendation features. And for good reason. Consider that in just the first quarter of 2018, Google reported that US$26.6-billion – or roughly 85 per cent of its quarterly revenue – came from advertising alone.
Twitter, a social-media giant that has traditionally owed its outsized stature to influence rather than revenue, brought in a profit for only the second time in its 10-year history after it increased advertising options and pumped up its sales strategy.
As a researcher in this area, it has been fascinating to witness the growth of our ideas and algorithms as they begin to affect our daily lives. But we don’t want to look back and say we missed an opportunity to transfer the lessons learned from the media and entertainment space into areas of social good.
Perhaps the most obvious space to begin with is health care. At present, global health-care spending is pegged at US$8.7-trillion, half of which is predicted to be spent on the three leading causes of death alone: heart disease, cancer and respiratory illness. The number of senior citizens is also expected to balloon to 604 million people over the next two years.
Imagine where that money could go instead if we were able to determine how to best allocate resources and perhaps even begin to eradicate some of these mass killers altogether with improved preventive care and drug treatments?
Now that AI has taken its nascent steps toward practical utility, we need to take a more mature view regarding where we can best deploy it. The AI research field needs a transition to a serious discussion regarding a refocusing of intellectual and financial investment that is cognizant of broader effects across our economy and society.
It’s therefore crucial that the AI talent making their way to industry also consider underrepresented and impactful industrial fields. A recent report from Element AI estimates the worldwide number of AI experts at a mere 22,000 people, many of whom have already been scooped up by the titans of tech such as Google, Microsoft and IBM. This top-heavy distribution is creating a mismatch between their companies and the industries that are ripe for AI augmentation.
So when I was considering my options for where to focus on this next phase of innovation, it was clear that the greatest needs arise in “traditional” areas of our economy: transportation, health care and finance. I chose finance because of the pervasive way that it affects our lives and the opportunity to apply image and video data – a largely untapped resource in financial services – to solve wide-ranging and impactful problems.
Until now, I’ve dedicated my career to computer vision, which is a field of AI that trains computers to “see,” process and understand the visual world. With my students at Simon Fraser University, we’ve been building algorithms that automatically interpret the world around them and make predictions about future events. For years, we applied this technology with great success to video analysis, sports analytics, and security and safety monitoring.
Visual data convey massive amounts of information that can be leveraged to analyze the evolution of our communities, land use and water supply, or better understand risk and disaster response. My team recently developed algorithms that analyze real-world satellite photographs taken after natural disasters and can effectively assist in rescue and reconstruction plans.
When major storms or earthquakes hit vulnerable areas, devastation can be minimized by automatically assessing damage based on AI-powered visual data and allocating resources much faster and more efficiently to the hardest-hit populations.
Complex relationships that combine imagery with multiple other data sources allow us to monitor patterns of how our world is evolving over time and make systematic predictions about the financial future of our neighbourhoods, our cities and our country.
Imagine what this could mean for areas as broad as the housing market to climate change to the distribution of natural resources. This can only happen if we focus our own resources – both intellectual and computational – on solving the right kind of AI problems. My hope is that more researchers will join me in diverting the talent flow to these vital but still underserved areas.