When someone I know applied for jobs after moving to a new country, they noticed something strange: despite having strong qualifications, they weren't getting interviews. Their CV looked perfect to human eyes, but something was filtering them out before anyone even saw it.

That "something" is likely an AI hiring system: software that screens CVs before they reach human recruiters. These systems are supposed to make hiring more efficient, but they can also reinforce bias and lock people out of opportunities, especially migrants and applicants from regions that aren't considered "prestigious."

How Automated Hiring Works

Most large companies now use Applicant Tracking Systems (ATS) that rely on AI to process CVs. Here's the typical flow:

1. Parsing: The AI extracts information from your CV: education, experience, skills, keywords.
2. Ranking: The system scores your CV based on how well it matches the job description and what the AI considers "ideal" qualifications.
3. Filtering: CVs below a certain score are automatically rejected.
4. Interview Selection: Only top-ranked CVs are sent to human recruiters for review.

The problem? These systems learn from historical hiring data, which often contains human biases. If a company historically hired more people from certain schools, regions, or backgrounds, the AI learns to prefer those same signals, even if they're not actually better qualifications.

Two Sample CVs: Equal Skills, Different Signals

To understand how bias can creep in, I created two sample CVs with equivalent skills but different "signals" that AI systems might interpret differently:

CV A: "Local Prestige" Signals

• Education: University of Oxford, UK
• Address: London, UK
• Previous employers: Well-known UK companies with recognizable names
• Job titles: Standard Western corporate titles (e.g., "Marketing Manager")
• Formatting: Clean, ATS-friendly format with standard sections
• Keywords: Matches common job description language

CV B: "Migrant/Foreign" Signals

• Education: Excellent university from another country (equally prestigious but less recognized)
• Address: Recent relocation, foreign address format
• Previous employers: Strong companies but with names that might not be familiar to the AI
• Job titles: Equivalent roles but with slightly different terminology or formatting
• Formatting: Slightly different structure (common in other countries' CV formats)
• Keywords: Same skills but expressed differently due to language/cultural differences

Both candidates have the same skills and experience, but CV A is more likely to pass through AI screening because it matches patterns the system has seen before and recognizes as "successful."

What AI Systems Might Overweight

Based on research and interviews, here are features that AI hiring systems often prioritize, sometimes unfairly:

Keywords

AI systems heavily weight specific keywords from job descriptions. If your CV uses slightly different terminology (common when translating experience across cultures), you might score lower even with equivalent qualifications.

Formatting

ATS systems are optimized for certain CV formats. CVs from other countries often use different structures, which can cause the AI to misparse information or miss important details.

"Western" Job Titles

Job titles vary across cultures. A "Project Coordinator" in one country might be called something else in another, but the AI might not recognize the equivalence.

School Recognition

AI systems trained on data from certain regions might not recognize excellent universities from other countries, even if they're equally prestigious. This disadvantages international applicants.

Address and Location

Some systems might filter based on location, which can disadvantage recent migrants or people applying from abroad, even if they're willing to relocate.

Real Experiences

I interviewed an HR professional who explained: "We use an ATS system, and honestly, I'm not sure how it's making some of its decisions. I've seen great candidates get filtered out, and when I look at their CVs manually, they're actually perfect for the role. But by the time I see them, if I see them at all, the process has moved on."

An immigrant student I spoke with shared: "I've applied to over 50 jobs. I have a master's degree and good experience, but I'm not getting interviews. My friends who are from here, with similar qualifications, are getting callbacks. I think the system doesn't recognize my university or my previous job titles."

What Job Seekers Can Do

While the system shouldn't require workarounds, here are practical steps applicants can take:

• Optimize keywords: Use exact terms from the job description in your CV.
• Use ATS-friendly formatting: Simple, clean layouts with standard sections (Education, Experience, Skills).
• Explain context: If you have international experience, briefly explain the equivalence (e.g., "equivalent to Marketing Manager").
• Network: Personal connections can help your CV bypass initial AI screening.
• Apply directly: Some companies allow direct applications that might skip ATS filtering.

What Employers Should Do

Companies using AI hiring tools have a responsibility to ensure fairness:

Policy-Level Changes

• Audit systems regularly: Check if certain groups are being filtered out disproportionately.
• Diversify training data: Ensure AI systems learn from diverse hiring examples, not just historical patterns.
• Human oversight: Always have humans review AI decisions, especially for borderline cases.
• Transparency: Be clear about using AI in hiring and allow candidates to request human review.
• Remove biased features: Don't let systems filter based on location, name, or other protected characteristics.

Practical Steps

• Test systems with diverse CV samples to identify bias.
• Provide training to recruiters on recognizing and countering AI bias.
• Create alternative application pathways for candidates who might be disadvantaged by AI screening.
• Regularly review and adjust AI scoring criteria based on actual job performance, not just CV signals.

The Global Impact

This isn't just about individual job seekers. It's about global mobility and opportunity. In an increasingly connected world, people should be able to work across borders based on their skills, not their origin. AI hiring systems that discriminate against migrants or international applicants create barriers to global talent flow.

This affects not just individuals, but entire economies. Companies miss out on talented candidates, and countries lose the benefits of diverse, skilled workers. The bias becomes self-reinforcing: if certain groups can't get jobs, they can't build the "recognizable" experience that AI systems prefer, creating a cycle of exclusion.