Women 'risk being left behind' in AI push, advocates warn Canadian lawmakers
asserted
advocates → risk → lawmakers
Parliamentary committee hears from experts on AI's impact on women, girls and 2SLGBTQ+ individuals
Canada must ensure its artificial intelligence strategy doesn't leave women and other vulnerable groups behind, advocates told lawmakers on Monday, amid concerns around AI's impact on historically disadvantaged communities and potential to perpetuate stereotypes.
asserted
advocates → hear → stereotypes
MPs heard from several experts during a House of Commons status of women committee meeting examining the impact of AI on women, girls and 2SLGBTQ+ individuals.
asserted
MPs → hear → women
"We stand at a pivotal moment," said Indra Kubicek, CEO of Digital Moment, a Montreal-based charity that works on digital literacy, particularly among girls and under-represented groups.
asserted
that → stand → girls
"My concern is that women in historically under-represented communities risk being left behind not because they lack ability, but because we are not creating the conditions for their success."
asserted
we → represent → success
Kubicek said women from diverse backgrounds as well as 2SLGBTQ+ individuals must participate in AI development and governance to avoid replicating biases.
asserted
women → say → biases
It learns from historical data, reflecting existing inequalities.
asserted
It → learn → inequalities
Without safeguards, we risk codifying those biases into systems influencing our lives," she said.
asserted
she → risk → lives
The committee is conducting a study on AI's effects on women, girls and 2SLGBTQ+ individuals, as well as what measures the Canadian government — which unveiled its "AI for All" national strategy in June — should take to prevent discrimination.
asserted
which → conduct → discrimination
As part of its $2-billion plan, the government said it would update laws and standards to "protect vulnerable groups from online violence and algorithmic biases."
asserted
it → say → violence
It said it would speed up AI's adoption in female-dominated sectors that are most vulnerable to disruptions, "unlocking new economic opportunities, softening transition shocks and confronting AI-fuelled online abuse that disproportionately targets women."
asserted
that → say → women
Prime Minister Mark Carney also signed a statement last month with other world leaders acknowledging the need to "strengthen oversight" amid "serious risks to safety and security" posed by advanced AI models.
asserted
Carney → sign → models
Women-dominated fields most affected
Human rights advocates and other observers have raised concerns over how AI systems will affect women and other vulnerable groups, including workers whose jobs could be replaced by the technology.
uncertain
jobs → dominate → technology
A March report from the United Nations' International Labour Organization (ILO) found that while generative AI could increase productivity and improve job quality, women-dominated industries were disproportionately affected.
uncertain
industries → find → quality
Around 29 per cent of female-dominated jobs are exposed to generative AI compared with 16 per cent of male-dominated ones, according to the report, which looked at occupations in more than 80 countries, not including Canada.
uncertain
which → dominate → Canada
Women were also over-represented in jobs at high risk of being replaced by AI automation, such as clerical and administrative roles.
asserted
Women → represent → roles
Separately, the UN Committee on the Elimination of Discrimination against Women (CEDAW) warned last week that AI systems "may reproduce and amplify gender stereotypes" and discrimination against women.
uncertain
systems → warn → women
"This is of particular concern where AI is used in areas affecting women's rights," such as employment, education and health care, the committee said.
asserted
committee → use → employment
More data needed
Amy Flood, executive director of Women's Health Collective Canada, said Monday that while AI presents a "very extraordinary opportunity" in health care, it also "risks making already invisible populations even more invisible" if it's developed with incomplete or biased data.
asserted
it → need → data
Flood noted that women have been under-represented in health research for decades, leading to "persistent gaps" in how reproductive health and other issues that disproportionately impact women are understood.
asserted
that → note → women
"AI is only as good as the data it's built on.
asserted
it → build → data
When that data is incomplete or biased, AI can reinforce existing [in]equities, leading to poorer health outcomes for women and gender diverse populations," she said during the committee meeting.
asserted
she → reinforce → meeting
She urged the federal government to invest in a national data strategy.
asserted
She → urge → strategy
"We can't build equitable AI systems using data that fails to adequately represent women," she said.
asserted
she → build → women
The Aporia
