Artificial intelligence has applications in practically every area of business, but many people are reluctant to try it because they don't trust it, says AltaML's Nicole Janssen.
That poses a problem for anybody developing AI solutions, suggests the co-founder and co-CEO of the Edmonton-based applied AI company.
"AI is seen as this black box thing that's taking people's jobs, and it's building killer robots. That's what the media and Hollywood has told us it is," Janssen told Taproot. "And so in order to change that mindset, you have to have this transparency with that end user as to why they're being told to do this, why this decision is justified, why this prediction is valid. Because that will build that trust that's needed in order to actually adopt AI."
AltaML made the Deloitte Technology Fast 500 and Canada's Top Growing Companies after recording three-year revenue growth of 2,309%. The company has more than 400 use cases under its belt in fields as wide-ranging as medical diagnostics, investment strategy, wildfire detection, emissions control, and preventative maintenance. So it's not that her company can't find work.
But globally speaking, Janssen said, only about one-fifth of AI solutions get operationalized, which leaves a lot on the table. "A big part of that is that lack of trust."
Building trust is core to the business case for responsible AI that she has been promoting in her role at the helm of AltaML.
Developing a diverse, ethical, and business-savvy workforce in AI is also key to her mission. AltaML's Talent Accelerators program has recruited, hired, and trained 200 data science interns to fill the tech talent gap. The contact with real-world situations is invaluable, said Janssen, who also sits on the board of Mitacs, an organization that connects the private sector and post-secondary institutions to solve problems.
"We're building great talent in Canada in academia. But when they get out, so much of that technical talent doesn't yet have that mindset of 'Is this valuable to business? Is this worth pursuing?' And so this internship allows them to get that experience."
Now she'd like to extend that paid internship opportunity to the less algorithmically inclined members of the team as well — the software developers, the project delivery managers, the product managers.
"AI doesn't work with just a team of data scientists," she said. "There's many people on each team that don't have the technical background ... That not only adds to the diversity of the team and the different experiences they come with, but it also comes with a little bit of a different mindset as to how they're looking at the problem and the project."
Janssen herself doesn't have a tech background, having built a career in event planning, real estate, and digital marketing before co-founding AltaML with her husband, Cory Janssen. They built the company in Edmonton largely due to the pipeline of data science talent coming out of the University of Alberta. Now they're expanding to Houston, with the northeastern U.S. to follow shortly. And they're raising a Series B round, so they can fund more ideas on the venture side of their business.
That level of activity from AltaML, the Alberta Machine Intelligence Institute (Amii), and others has made Edmonton one of Canada's leaders in AI. Canada, in turn, is one of the top countries in the world investing in this area. Leaning into responsible AI is an opportunity for Canada to differentiate itself from much bigger players, Janssen said.
"Canada comes to the table with an existing reputation of being responsible," she said. "So why can we not carry that through into AI, and be the leader in responsible AI globally?"
Achieving that requires governments, academia, and industry to work together so that ethics is baked into the commercialization of AI and machine learning, she said.
"We're on our way, but let's really shoot for the stars," she said.
Katrina Ingram of Ethically Aligned AI agreed that Canada is well-positioned to be a leader in this space, but it needs to keep its foot on the gas.
"Canada had a lot of early wins," she said, citing the Algorithmic Impact Assessment tool developed by the Treasury Board and the Pan-Canadian Artificial Intelligence Strategy.
"But what I've seen lately is that we are falling a little bit behind," noting that Bill C-27 is still grinding its way through Parliament, while Europe's GDPR is nearly five years old.
For Ingram, regulation is a necessary piece of the puzzle when it comes to ensuring AI is trustworthy and responsible. While leaders like Janssen can set the tone from the top and structure their teams to ensure professional responsibility, it's important to have rules that everyone must follow, even if doing so stands in the way of profit. Otherwise, "you create an uneven playing field for companies who are maybe taking a hit by doing the right things while other companies are just doing whatever they wish."
Hear more from Janssen and Ingram on Episode 39 of Bloom, Taproot's podcast about innovation in Edmonton.