AI no solution for immigrant job seekers, study finds

By Diary Marif
Local Journalism Initiative Reporter

Artificial intelligence tools currently on the market to help immigrants navigate job searches are far too inadequate for the challenges facing newcomers, a new study has found.

Academic researchers undertook a year-long study to assess the usefulness of these AI-enabled tools. They found that AI tools, though useful for entry-level job seekers, do not reflect the real barriers new immigrants face when entering the job market. AI tools were primarily designed for practical tasks like job searching, resume building, and interview preparation.

Labour market integration for immigrants, on the other hand, requires navigating complex issues such as licensing requirements, weak professional networks, and unequal access to opportunities, which remain unaccounted for in these AI tools. The researchers are pushing for more “human-centered technology” – as opposed to efficiency-focused tools – which they say can be achieved through greater collaboration with developers, settlement practitioners, and immigrant communities.

“Advancing technology alone cannot fully address systemic barriers facing skilled immigrants,” the researcher said in email comments to New Canadian Media.

The Government of Canada cautions hiring managers that “AI can overlook cultural nuances” and suggests working with diversity and inclusion specialists as part of the hiring process. The use of AI has become common practice on both sides of the hiring process. It is being used actively to screen applicants by searching for key terms in resumes to help optimize the hiring process.

The researchers – a university professor and two immigrant settlement experts – observed that AI tools fail to take into account broader aspects of immigrant labour market integration like recognizing transferable skills and non-linear career trajectories. The researchers also emphasize the need for sustainable funding models to update older models, maintain them and make them more realistic.

This study was part of one of the five panels presented at a migration and integration conference at UBC on April 27.

This study was titled “Evaluating the Effectiveness of Advanced Digital Technologies to Support Immigrant Professionals’ Integration into the Canadian Labour Market.” and was led by Dr. Anusha Kassan, UBC associate professor of Educational and Counselling Psychology, and presented by Fatemeh Kazemi, senior researcher at the Centre for Immigrant Research at The Immigrant Education Society (TIES) and Kreisa Hilaro, manager at TIES Centre for Immigrant Research.

“Based on our experience working in the sector, we are already seeing growing gaps in service capacity and support accessibility,” the team said. “This is especially concerning because newcomers often rely heavily on these services during their first years in Canada.”

The conference also explored questions of identity, integration, and political behaviour. Aodi Cheng, an MA student in political science at UBC, examined how digital platforms and information environments shape attitudes toward immigration and integration among Chinese Canadians. Maxime Coulombe, Affiliate Assistant Professor at Concordia University, studied public opinion on government use of artificial intelligence, focusing on support, bias, and political behaviour using survey and experimental methods.

Irene Bloemraad, one of the conference organizers, told NCM that this year’s conference focused on impact beyond academic circles. She said migration scholars have long tried to inform public debate with evidence, but a gap still exists between university research and broader public understanding.

Bloemraad also said that researchers are increasingly using technology in their work. Several projects went beyond traditional methods, using tools such as AI-generated images and interactive storytelling in outdoor spaces, to explore how people make sense of complex issues like artificial intelligence and the sense of belonging.

Ottawa’s warning on AI in hiring

Artificial intelligence may help speed up hiring, but federal guidance warns that it can also deepen unfairness if employers use it without human judgment, clear rules and proper oversight.

The Public Service Commission of Canada says hiring managers remain accountable for every decision made in an appointment process, even when AI tools are used to draft job ads, screen applications, prepare interview questions or assess candidates. Managers must understand the AI systems they use, be able to explain their decisions, and validate AI-generated material for accuracy, relevance, bias and barriers.

The warning is especially relevant to immigrant job seekers, who may already face barriers around foreign credentials, language, cultural expectations, professional networks and non-linear career paths.

Key points from the federal guide

Human oversight is required
Hiring managers cannot hand responsibility to an algorithm. They must review and adapt AI-generated material to reflect the actual job and remove or reduce bias.

AI can miss cultural context
The guide warns that AI can overlook “contextual and cultural nuances,” especially in communication with candidates. It recommends review by hiring managers and diversity and inclusion specialists to make sure AI-generated messages are appropriate and fair.

Keyword screening carries risks
The commission gives an example of inappropriate AI use: a tool that ranks resumés using unclear criteria such as keywords or biased historical data. Such systems can produce scores that recruiters may rely on without understanding how they were created.

Candidates must be told when AI is used
Before assessments, candidates should be informed if AI tools will be used. The guide says transparency is essential because it helps build confidence in the fairness and integrity of the process.

AI scoring must be explainable
AI tools should only be used to assess candidates if hiring managers can clearly explain the role the tool played, what criteria or data it used, what assessment it produced, and how that information shaped the final decision.

Translation errors can disadvantage applicants
The commission says information shared in English and French must be the same throughout the hiring process because AI tools can produce translation errors. That matters in federal hiring, where candidates should not be disadvantaged by the official language they use.

Rules for candidate use of AI must be clear
Employers should tell applicants whether they can use AI to prepare cover letters, answer screening questions or complete assessments. If AI is allowed, candidates should be told what tools they may use and when disclosure is required.

AI access is unequal
The guide notes that candidates may not have equal access to paid AI tools, strong language support, technical knowledge or proper computer equipment. That gap can affect how applicants perform in AI-influenced hiring processes.

AI-detection tools are discouraged
The federal guide warns against using tools that claim to detect whether a candidate used AI. It says they are unreliable and can produce false positives with serious consequences for applicants. Reassessment or follow-up questioning is safer than relying on detection software.

Bias checks are mandatory
Since July 1, 2023, federal hiring rules require assessment methods to be reviewed for biases or barriers that could disadvantage people from equity-seeking groups. If problems are found, reasonable efforts must be made to remove or reduce them.

Why it matters

The federal guidance supports the study’s central finding: AI can help with simple hiring tasks, but it cannot solve the deeper obstacles newcomers face in Canada’s labour market.

For immigrant professionals, the danger is that AI hiring systems may reward polished keywords while missing international experience, transferable skills, interrupted careers, foreign credentials and culturally different ways of presenting qualifications. That makes human judgment, settlement-sector expertise and transparent hiring rules essential.