Skip to main content
newsletter

Artificial intelligence (AI) can be a powerful tool to help build a more inclusive economy.ljubaphoto

It’s no secret that the pandemic resulted in women and marginalized communities being ousted from the work force in record numbers. Though many demographic sectors have since bounced back, the gains remain unequal among traditionally under-represented groups.

For example, employment in the accommodation and food service industries, which are traditionally staffed primarily by women, are still 17 per cent below pre-pandemic levels. And while the unemployment rate for racialized workers has returned to pre-pandemic levels, it’s still higher than that of non-racialized workers. Youth, Indigenous people, women with children under 6 and many other vulnerable sectors have experienced similar disparities.

“We’ve gotten back to a pre-pandemic level and this is something to celebrate, but is it good enough? If we want an inclusive economy, we have to put in the effort,” Kaylie Tiessen, economist and policy analyst at Unifor, said recently to The Canadian Press.

The question is: How can we raise these under-represented sectors to an equitable rate of participation in the labour market?

Artificial intelligence, or AI, might be the answer.

Designing AI ‘in accordance with our values’

“[AI] could potentially play a beneficial role in addressing Canada’s overall hiring crisis,” says Dr. Marianna Ganapini, faculty director and instructional designer at The Montreal AI Ethics Institute.

Dr. Ganapini acknowledges that algorithms can “assimilate and reproduce the unjust biases present in our society.”

There are well-known examples, like a recruiting engine at Amazon launched in 2014 that was fed by machine learning and produced results that were reportedly biased against women candidates. (The shopping giant discarded the program in 2017.)

But Dr. Ganapini notes that it’s possible to counteract this potential harm by leading with a “value sensitive design strategy,” which lays out “actional steps for designing technology in accordance with our values.”

A value-sensitive design process lets stakeholders figure out what values they want to embed in a system, design that AI system in accordance with those values, and then find ways to check whether the system behaves ethically, Dr. Ganapini says.

Identifying pain points

Diversio Global, based in Toronto, has developed an ethically stewarded AI-based platform that enables companies to measure and then improve diversity, equity and inclusion (DEI) in their organizations.

First, the platform collects data through an anonymous survey which asks employees about their ethnicity, gender, sexual orientation, disability and experiences with bias or workplace harassment.

The platform then combines that data with other information about the company to discover what equitable representation looks like considering the nuanced experiences of unique groups.

Once this information has been amassed, the platform uses algorithms to identify “pain points” within the organization, uncover bias and suggest actionable solutions, both for the company as a whole and within specific departments or roles.

“When we give folks the data to lay as their foundation to begin the work of DEI, it becomes a launch pad in which I’m able to work with them to build trust and strategic tactics,” says Brooke Graham, DEI consultant for Diversio.

One tool among many

In one recent case, Diversio worked with an organization losing female employees at a disproportionate rate compared to male counterparts. After collecting and analyzing information through their platform, Diversio was able to identify the organization’s primary pain point: that women felt there was a lack of interest and investment in their professional development.

Diversio helped implement a mentorship program that identified high-potential and high-performing employees from under-represented backgrounds to pair with high-tier mentors. They also provided training and other support to ensure every participant benefited as much as possible from the experience.

It worked; Ms. Graham says that one year after consulting with Diversio, “two times more women advanced to leadership in their most recent promotion cycle, and gender diversity increased in the organization by 40 per cent.”

There’s no silver bullet that will solve the problems of bias and underrepresentation in the workplace. Both Dr. Ganapini and Ms. Graham agree that AI is just one tool among many required to correct the inequity in Canada’s labour market – reforms such as accessible child care, affordable housing and anti-racism initiatives are also critical.

But more implementation of ethical AI could be an important step to help the tens of thousands of Canadians facing barriers to employment, empowering organizations to face their DEI shortcomings head-on.

Ask Women and Work

Have a question about your work life? E-mail us at GWC@globeandmail.com.

Question: I’m on mat leave and heading back to work in a month. However, I’ve just been told by my supervisor that I will need to work 10 a.m. to 7 p.m. when I return. Before I had my baby, I worked until 6 p.m., sometimes 7, but now I need to get to my daycare by 6 p.m. each day (which means leaving around 5:30 each day). My supervisor’s rationale is that we work a lot with people on the West Coast, but I feel like this is just an excuse to get me to quit (or fire me) because I’m now a mom and won’t be at their beck and call. Can I fight this?

We asked Simone Ostrowski, partner at Toronto-based Whitten and Lublin Employment Lawyers, to field this one:

Yes, you can fight this. The Ontario Human Rights Code provides that employees are entitled to accommodation at work based on their family status. ‘Family status’ in this context includes only significant parent-child responsibilities, such as child care and eldercare. Workplace accommodations can include working a specific schedule, having flexible work hours or duties and working from home. The employee also has a duty to co-operate in the accommodation process which means being flexible, pro-active and willing to compromise in finding a solution.

For an employer to refuse to accommodate an employee based on undue hardship, they must prove that the accommodation would be unfeasible based on cost and/or health and safety considerations, and not merely inconvenient or bad for morale.

The central question is whether you can perform the essential functions of your role on a work schedule that allows you to leave at 5:30 p.m. each day. Some issues to consider are:

  1. Could a family member handle daycare pickup for you, allowing you to work until 7 p.m. some or all of the time? Why or why not? You should always be prepared to explain why you specifically must leave work to do the pickup.
  2. Could you effectively address any issues that arise from the West Coast or otherwise after 5:30 p.m. the following workday, or are these urgent issues that must be attended to on the spot?
  3. If you must deal with certain work issues arising after 5:30 p.m. immediately, could a colleague handle these issues in your place? It may be that only one employee must stay until 7 p.m. for coverage purposes, which does not need to be you.
  4. If the work issues arising after 5:30 p.m. must be dealt with urgently, could you work an hour or two in the evenings from home after picking up your child and/or after they go to sleep?

These are only some initial questions to consider when assessing whether your employer has a duty to accommodate your new scheduling needs. Fortunately, as we have learned over the pandemic, many employees can perform their jobs just fine under flexible work schedules and remote work arrangements.

Interested in more perspectives about women in the workplace? Find all stories on the hub here, and subscribe to the new Women and Work newsletter here. Have feedback on the series? E-mail us at GWC@globeandmail.com.