Still not convinced about AI?

When you think of a robot, do you picture a benevolent one? Probably not, according to Terry Hickey, chief analytics o•cer at CIBC in Toronto. Movies very often portray artificial intelligence (AI) and robots in a bad light.

There’s one Wall-E for every 10 Terminators, he says.

“We have a perception problem. Hollywood has told us that AI is bad, it’s evil, it’s coming after our jobs. And we should lock it in a closet.”

But organizations need to be thinking differently about AI and what it’s going to do for employees, says Hickey, speaking at the Human Resources Professionals Association (HRPA) conference in January.

“It’s technology that allows us to replicate human capabilities through a machine that has the ability to learn and, most importantly, to augment our employees… It’s going to change everything that you do, not just the people that you support.”

But it’s not perfect, he says.

“It’s still new, we’re still training it, and it’s still got to figure a bunch of things out. It’s going to learn over time. And, eventually, it may get this right. But this technology is still in its infancy…. Although AI seems to be the word that we’re talking about all the time, it’s really relatively new in business.”

While people have heard about the benefits of AI — such as greater efficiencies, reduced operational costs and faster decision-making — there are also limitations, such as a narrow focus, ethical issues and bias, says Hickey. But employers and HR should definitely be getting onboard, and there are a few basic steps for getting started.

Tips on bringing in AI

For one, consider the impact of AI. At CIBC, that has meant focusing on small projects that would take somewhere between four and seven weeks to accomplish, instead of multi-million-dollar projects that could have a greater impact but also greater risk, he says.

Another area of focus is transparency. If employees are not told about what AI means to the organization and how they should be thinking about it, then they’re going to have that negative perception enforced by Hollywood, says Hickey.

“It was almost six, seven months that I was in the organization before I was even allowed to say the words AI. 

Because everyone was afraid: ‘Terry, as soon as you say those words, everyone’s going to get worried that their job’s at risk.’ So, we came up with a communications plan that allows us to educate the organization [about] what it is and how they need to be thinking about it, and how we’re going to use it to augment the people within our organization.” 

Co-ordination is another key consideration, says Hickey, as a large organization like CIBC couldn’t have 15 different groups going out and trying AI willy-nilly.

“We set up a central team that basically organized how we’re going to do it, what tools that we’re going to use, what methodology that we’re going to use. So that’s co-ordinated [and] meant we’re going to do this together. It may mean we do it a little bit slower, but you can’t run the risk of doing this wrong; you have to do this right.”

As for how AI can be used exactly, HR should be thinking about issues such as recruitment, where technology looks at how people behave during video interviews, or engagement and retention, where the technology looks at internal data to identify high-potentials or people likely to leave, he says.

It's also important not to let your organization become too dependent on AI, he says. 

“The people in your organization still need the skills, the underlying skills of what it is that’s being automated, in the event that that technology isn’t available in the future.”

HR’s role in evolving AI at work

For HR, AI is a golden opportunity to think about the work that happens at their organization by answering important questions, says Hickey.

“What work should be done by your employees or your humans? What work should be done by contractors or flexible workers? What work should be done by professional services? [Do] we need to rent that skill because we don’t have it or do we need to augment our team for a certain period of time? What work should we outsource? If it’s not critical to our business, should we even be in that space?”

And, most importantly, what work can now be done through automation and artificial intelligence?

“You have the opportunity to set that stage today. And you need to lead the organizations when new work comes up or even with existing work. We need to start squatting that work into the appropriate bucket. You need to start creating that backlog of opportunity for AI within your team,” he says. 

Once the strategy is figured out, the hardest part is educating the workforce. 

“In a smaller organization, you could probably do that in an afternoon or a week. In a large organization, it took us almost six months to go around the organization and tell everyone what is it that we’re doing, where we’re going, giving people the opportunity to talk to us about what they want changed in the plans that we were developing,” says Hickey.

It’s also about talking about the technologies that are going to be used and what skills and training are required, he says. 

“You need to start working with people so that they understand the impacts. And you need to start to develop them so they can get the skills that they need to be able to live in this new economy.” 

HR also needs to develop use cases, he says. At CIBC, that meant setting up meetings with every pocket of the organization, such as finance or retail banking, to find out where AI could help, says Hickey.

“We created a backlog of 160 use cases within our organization to say, ‘This is what we’re going to do. This becomes the work that we’re going to create for the next year or two or three within the organization…’ In our first year, we got through about 42.”

And every one of the use cases should have a business case, he says, thinking about: Am I going to make more money? Is it going to cost me less than to run my business more efficiently? Or is it going to improve my risk profile?

And while HR may be familiar with using tech in areas such as payroll software or an HRIS, it could be a challenge picking an appropriate vendor in the AI space.

“If you're not developing your own solutions and you're buying a thirdparty tool or AI solution, how do you know that they've trained it the right way? How do you know that they didn't train it to be biased against women or biased against minorities?” says Hickey.

“You actually have to take some incremental steps in testing the solutions to make sure that those biases don't exist because, at the end of the day, your organization is the one that's at risk here, your customers are going to sue you — they don't even know about this third party in the background. So… people need to be careful about who they go to.”

That means sitting down with vendors, he says, to truly understand who their clients are, how their data is protected and how have they tried to minimize bias within their solution.

Developing a talent pipeline

Another important action item is developing a new talent pipeline, says Hickey. That means finding out where the workers are going to come from to carry out the AI.

“For the most part, a lot of this talent doesn’t necessarily exist within your organization. These people are still in school or they’re just recently out of school. So [it’s about] how can you set up arrangements with universities or colleges to find that talent today, so that they can come into your organization over the next couple of years to help you make a difference. Establish those partnerships.”

Overall, when it comes to newer AI tools, HR may not be fully ready yet for this role, he says. While there are certain groups within organizations that are more advanced in the adoption of AI, it’s usually the more technical people who are going to figure out how it can be deployed.

“That’s where HR is going to become critical… [to determine]: How do we scale that across the organization?,” says Hickey. “We have to rely on HR to help us change job descriptions, we have to rely on HR to convince people that this technology isn’t out to get your job. So, I think HR plays a critical role. I don’t think that we’re there yet. I think that it’s probably going to take us another year or two to get to the point where HR is leading the charge to get people ready.”

GLOBALLY, HR SEES LOWER COSTS WITH AI

• 81% Share of workforce expected to be retrained in next three years due to AI adoption
• 57% Share of workforce retrained in past year because of AI adoption
• 56% Average revenue increase in HR function adopting AI
• 55% Average cost decrease in HR function adopting AI

Source: McKinsey

AI BENEFITS APPRECIATED AT WORK

Number of workers who would embrace AI if it…

• 64% simplified or automated time-consuming internal processes
• 64% helped better balance their workload
• 62% increased fairness in subjective decisions
• 57% ensured managers made better choices affecting employees

Source: The Workforce Institute, Kronos