Canada has fifth-biggest AI workforce, but still lacks diversity: study

TORONTO Canada has the globe’s fifth-largest artificial intelligence workforce, but is still far from closing the gender gap in the sector, according to new rankings from the World Economic Forum.

The international organization used Linkedin data to find the size of Canada’s AI workforce is lagging behind the U.S., India, Germany and Switzerland, but beating dozens of countries including France, Spain, Singapore and Sweden.

However, the Forum found as part of its annual rankings on gender disparities that women only make up 24 per cent of Canada’s AI workforce, and 22 per cent of the world’s AI workforce.

The Forum said the lack of women in Canada’s and the global AI pool is troubling because it implies that technology is being developed without diverse talent, thus “limiting its innovation and inclusive capacity.”

Low integration of women in AI talent pools, it said, is a “significant missed opportunity in a professional domain where there is already insufficient supply of adequately qualified labour.”

If not addressed soon, the Forum warned the gap could widen further.

Sarah Kaplan, the director of the University of Toronto’s Institute for Gender and the Economy, said Canada needs to do more to develop the diversity of the AI workforce because the technology has the potential to replace lots of human processes and decisions.

“If we don’t have a diverse workforce working in AI, we risk not only perpetuating existing biases, but actually amplifying them and leading to really negative outcomes for the most vulnerable people in our society,” she said. “We need to make a really huge deal out of the fact that we are not bringing a diverse workforce to work on this technology, which may shape all of our lives.”

Diversity, she added, is important because whoever designs AI technologies will decide how the systems act.

A wider talent pool can help mitigate situations society has already encountered, such as where some facial recognition systems were developed without the ability to recognize black faces.

“No one ever thought ‘does this work on someone other than a Caucasian person?”’ Kaplan said. “The fact that we have an AI technology developed by a non-diverse workforce is hugely problematic.”

Kaplan said she has grown tired of seeing AI firms treat diversity as a “afterthought” and wishes it was being taken more seriously, so Canada’s gender gap would be much better than the global average’s.

The Forum did find at least one positive: Canada’s AI sector employs more women than the four countries — the U.S., India, Germany and Switzerland — that rank above Canada in AI workforce size.