Navigating the use of artificial intelligence in Canada’s newcomer settlement

From October 8 to 10, 2024, the Ontario Council of Agencies Serving Immigrants Leaders’ Forum gathered newcomer settlement sector leaders to discuss public policy, programs and services; best practices on governance and new trends. About 150 people, including funders from all levels of government, attended the multi-day event at the Pan Pacific Hotel in Toronto, Ont. It included panel discussions on topics like gender-based violence in the workplace, enhancing legal support for migrant workers and opportunities for artificial intelligence (AI) within newcomer settlement.

Wendy Cukier, academic director and founder of the Diversity Institute (DI) and academic research director of the Future Skills Centre (FSC), participated in a panel on Oct. 9 titled, AI and Newcomer Settlement: Navigating Opportunities and Challenges in Canada. Moderated by Marco Campana, knowledge mobilization and social action coordinator at Access Alliance Multicultural Health and Community Services, panellists discussed the strengths and weaknesses of AI use in the settlement sector and ethical considerations around its implementation. 

Darcy MacCallum, CEO of Social Enterprise for Canada, highlighted his organization's focus on foreign credential recognition and how they’re narrowing down the pain points in the system to try to find solutions where AI can help. “We have to be cautious and curious when it comes to AI. We need a process of curiosity together because its use is inevitable,” MacCallum said. Jennifer Freeman, CEO of PeaceGeeks, spoke about how AI can support integrating newcomers into the labour market. She highlighted a tool they developed to help newcomers navigate the economy with roles, interview skills and recruitment preparation. 

Petra Molnar, associate director of the Refugee Law Lab and professor at the Osgoode Hall Law School and the Centre for Refugee Studies at York University, highlighted the risks of AI use and spoke about how technology exacerbates power differences in society. Focusing on migration and how AI is used for surveillance at borders and refugee camps, she said security is prioritized over providing basic needs like water, shelter and humane living conditions. “Technology is being implemented at the expense of human rights,” Molnar said. 

Cukier opened by highlighting the importance of understanding the role of AI for job seekers. Recent research by DI, the Environics Institute for Survey Research and FSC has highlighted that while new jobs will emerge and some old jobs will disappear, virtually all jobs will be transformed by technology. Agencies focused on supporting pathways for newcomers to employment need to ensure that they recognize the importance of AI skills. The Advanced Digital and Professional Training Program (ADaPT) and ADaPT for Black Youth, for example, are building out AI skills competency frameworks and courses. ADaPT is an employer-focused work-integration learning program with a proven track record at getting participants jobs. She stressed the opportunities that AI presents for organizations across sectors, including for settlement agencies at every stage of their value chains, whether in supporting clients, streamlining internal processes, or helping reduce the time and effort spent on proposal writing, reporting and support activities. Diversity Institute research has shown that regardless of what employers are doing, employees are using AI tools, and it's critical that organizations across sectors put in place the appropriate policies and processes to provide “guardrails.” 

“Many people using free versions of generative AI tools, for example, are unaware of the risks of errors and intellectual property violations,” Cukier said. But, as with other technologies, AI is a tool that can be used for good or for evil. There is lots of evidence of the ways in which it can exacerbate bias and inequality, she said. “Our aim is to have an EDI [equity, diversity and inclusion] lens with AI as it pertains to supporting refugees integrating into the labour market and acquiring necessary skills.” 

Cukier noted that AI can streamline processes at every stage of the settlement journey. Adaptive learning, particularly for language development skills, for example, has tremendous power, not just for newcomers but for pre-arrival training. “Facility in French or English is one of the key enablers of access to employment and advancement, but we do not do a good job of language training in Canada,” she said. “The Auditor General of Ontario, for example, said many people would learn more language driving a cab than they do sitting in classrooms. And the costs are considerable.” An area of focus for DI and FSC is on innovative ways to use technology for upskilling and reskilling around language training. 

While other panelists talked about the need for more investments by Immigration, Refugees and Citizenship Canada in the settlement sector to support technology acquisition, Cukier maintained the need for a “whole of government” approach to digital transformation. This approach recognizes the needs of non-profit as well as for-profit organizations with new investments in AI adoption and management. She highlighted the challenges of fragmentation and duplication in the sector and noted there have been a plethora of online services like job boards created that meant newcomers and employers had to go to multiple sources for support. 

Cukier advocated for more focus on high quality, well-resourced, shared platforms, like some of the services discussed by Freeman being developed by PeaceGeeks to support navigating occupational requirements, which can be white labeled. Building economies of scale for small settlement services providers who do not have the resources to develop technologies should be the focus so settlement agencies can “stick to their knitting” rather than trying to build out their IT departments and try to develop deep AI skills. Instead of building individual job boards, for example, the Lifeline Challenge (external link)  platform, Powered by Magnet (external link) , helps connect Afghan and Ukrainian newcomers of all backgrounds and levels of experience to jobs.

Cukier noted that the potential of AI to reinforce biases is a significant concern. Biases in AI systems arise when they are trained on historical data that reflects inequities, she said, and this can reinforce patterns of unfair treatment in areas like housing, employment and access to services. To minimize bias, AI systems for newcomers should be custom-built with newcomers’ specific needs in mind to address their unique challenges. Additionally, Cukier said frameworks provide a consistent approach, ensuring that AI is implemented ethically and effectively to meet the needs of newcomers, adding “There are lots of ways AI can help organizations become more productive, but it can't be at the expense of human rights.”