If you’ve ever applied for a job and heard nothing back, you’re not alone. These days, a lot of applications are screened by software before a human ever sees them. That can save employers time, but it can also create a new problem.
A Stanford HAI write-up of a large-scale US study suggests that some AI screening tools can produce unequal outcomes for different racial groups, and can even create “systemic rejection”, where the same person gets filtered out again and again across multiple applications.
Scope note (US content)
This research is based on US hiring data and uses US legal definitions (including the EEOC’s “four-fifths rule” and Title VII). The UK has different laws and guidance, so treat this as strong evidence of risk, not a direct measure of what’s happening in every UK hiring process.
What the research says (in plain English)
The study analysed 4 million applications from 3.4 million people, across 150 employers in 11 sectors, all screened through a single third-party vendor.
Here are the headline findings:
- Unequal outcomes showed up job-by-job. When the researchers checked each role separately, they found signs of “adverse impact”, meaning one group was recommended at a much lower rate than another.
- Black and Asian applicants were more likely to be affected. The Stanford HAI write-up reports that 26% of Black applicants and 15% of Asian applicants applied to roles where the AI system disadvantaged their group.
- Some people get shut out repeatedly. Of applicants who submitted four applications, 10% were rejected everywhere they applied, at a higher rate than you’d expect if each employer’s decision was truly independent.
Why this matters for your first chance
When you’re 16 to 25, you’re often applying without a long work history, without a polished CV, and sometimes without the “right” keywords.
If screening tools are trained on past hiring patterns, they can end up copying yesterday’s biases. If lots of employers use the same vendor, one system can quietly shape outcomes across a whole local market.
That is exactly why inyourroots® is building a strengths-first, skills-first route into work. Your first chance should not depend on whether you already know how to “sound employable” on paper.
- A clear human route in: SMEs get structure and support, but humans stay accountable for decisions, especially when the stakes are high.
- Bias-aware by design: we assume bias is likely, so we build with safeguards and monitoring in mind from day one, rather than retrofitting after harm.
- No monoculture scoring: we avoid one opaque pass or fail score being reused across multiple employers as a universal filter.
- Explainable matching: young people can see why something matched, and what to do next. Not just a hidden score and a rejection.
- Strengths-first, not pedigree-first: we are not rewarding polish, school brand, or confidence. We are surfacing potential, fit, and readiness.
If AI is going to sit upstream of opportunity, it cannot be a silent gatekeeper. The research is a useful warning, and it also gives us a clear build brief.
How inyourroots® is building a fairer route in
UK context (why it still matters here)
Even though the dataset is US-based, the UK is actively paying attention to automated decision-making in recruitment. The Information Commissioner’s Office (ICO) has published guidance for organisations and jobseekers on using automated tools fairly, lawfully, and transparently.
The takeaway is not “AI is bad”. It’s this: people deserve clarity, safeguards, and a fair shot, especially at entry level.
What you can do if you’re applying (practical, low-pressure)
You should not have to “game the system”, but these steps can help you get seen.
- 1.Apply with evidence, not just claims
- Add examples, mini projects, volunteering, short courses, or anything you’ve actually done.
- 2.Use the job description’s language (without copying it)
- If the role says “customer service”, make sure you use that phrase if it’s true for you.
- 3.Ask for feedback when you can
- A simple message like, “Is there anything I can improve for next time?” can unlock useful info.
- 4.Spread your chances across different routes
- Don’t rely only on one platform. Try local employers, community groups, school/college links, and direct applications.
What SMEs can do (if you’re hiring)
If you’re a local employer using screening tools, you can still protect fairness.
- Check outcomes per role, not just “overall”. Averages can hide problems.
- Keep a human review step for entry-level roles.
- Be transparent about what you use and what you look for.
- Offer an alternative route for applicants who struggle with standard forms.
The inyourroots® promise
We’re building a local ecosystem where young people can be seen for how they work best, what they can do, and what they’re ready to grow into, with no bias.
If you want a first chance that’s based on strengths, not polish, join the inyourroots® career discovery journey.
- Start your inyourroots® career discovery journey
- SME? Become a founding partner and help shape fair, local hiring