10-Minute Rejection? This is what is happening when CVs are instantly binned

I’ve been in recruitment long enough to know when something’s off. Over the past year, I’ve noticed a pattern that’s been troubling me. Qualified candidates with exactly the right experience, the right clearances, the right background for defence and engineering roles, are getting rejection emails within minutes of applying.

Not hours. Not days. Minutes.

At first, I thought it was just bad luck. But when dozens of candidates report the same story, you start asking questions.

What I found

It turns out I’m not imagining it. In November 2024, the UK’s Information Commissioner’s Office (ICO) published audit results on AI recruitment tools. The findings were bleak. They made 296 recommendations to AI providers for compliance failures.

Here’s what the ICO found:

AI systems were inadvertently categorising candidates by race, gender, and sexual orientation

Many lacked proper bias testing

Some were scraping excessive data from LinkedIn and social media

Providers were struggling to explain their own algorithms’ decisions

Yet 30% of UK employers now use AI in recruitment. Nearly one in three applications may never reach human eyes.

The first UK case

In 2024, Pa Edrissa Manjang, a black Uber Eats driver in Oxford, received a financial settlement after proving that Microsoft’s facial recognition software discriminated against him. The system repeatedly failed to recognise him, Microsoft admitted its software worked less effectively for ethnic minorities, removing him from the platform.

The Equality and Human Rights Commission funded his case. That’s significant. It means the UK’s equality watchdog is taking AI discrimination seriously enough to back claimants.

His lawyer, Hannah Wright, was clear that we need better protections and transparency in AI-based employment decisions.

Why defence sector Is particularly vulnerable

This matters especially in defence and engineering recruitment, and here’s why:

UK-specific qualifications aren’t recognised. Many AI recruitment systems are trained on US data. They don’t understand HNCs, UK apprenticeship pathways, or MOD contractor experience. Your perfectly relevant background might be filtered out because an algorithm doesn’t recognise British engineering qualifications.

Security clearance creates gaps. AI sees an employment gap during your SC clearance process and flags it negatively. It doesn’t understand that’s normal for cleared roles.

Career breaks get penalised. We already have too few women in defence engineering. AI trained to prefer “continuous employment” discriminates against anyone who has taken time off, for childcare, health, or any other legitimate reason. This perpetuates gender imbalance and potentially breaches the Equality Act 2010.

Regional concentration looks suspicious to algorithms. Defence work is concentrated, for example, in Portsmouth, Barrow, Bristol, and Scotland. AI using postcodes as proxies might discriminate against candidates from defence hubs. These are the very people most likely to understand the sector.

Technology remains unreliable. AI-generated search strings often produce inconsistent or unexpected results compared to manually crafted ones. If the technology struggles with search string logic, what happens when making binary decisions about candidate suitability?

What UK law says

Under the Equality Act 2010, an algorithm counts as a “provision, criterion, or practice.” If it discriminates, even unintentionally, employers are liable.

Compensation in UK Employment Tribunals is uncapped. The most serious discrimination cases can result in awards of around £50,000, covering both financial loss and injury to feelings.

The legal framework exists. The challenge is proving it, algorithms are “black boxes,” and getting vendors to explain their decisions is nearly impossible.

What this means

If you’re a candidate: That instant rejection might have nothing to do with your qualifications.

You’re not imagining it. The system is flawed, and according to research, 73% of UK candidates would be deterred from applying to a company if they knew AI was being used to screen applications. You’re not alone in wanting human judgement.

If you’re an employer: When did you last audit your recruitment tech for bias? The ICO made 296 recommendations. Is your vendor compliant? In defence procurement, where social value and diversity are important, can you afford the reputational and legal risks?

Why we do things differently

This isn’t a sales pitch, it’s a reality check. Even for experienced professionals, AI-generated approaches can produce unreliable results. At The Thrive Team, we’re headhunters. Every candidate we put forward has been personally assessed by someone who understands UK defence and engineering sectors.

We know which qualifications matter. We understand clearance timelines. We recognise transferable skills that algorithms miss. We can explain our decisions and be held accountable for them.

When I place someone, I’m putting my professional reputation behind that recommendation. An algorithm doesn’t have a reputation to lose.

The irony is that AI was supposed to reduce bias in recruitment. The ICO’s November 2024 investigation shows it’s often doing the opposite.

67% of UK job applicants think AI lacks the nuances of human judgement. They’re right.

In a sector where you’re looking for specialised skills, security clearance eligibility, and the ability to work on sensitive projects, you can’t afford to let an algorithm make the call.

Have you experienced instant rejection despite being qualified? I’d like to hear about it. Connect with me on LinkedIn let’s have an honest conversation about what’s happening in defence recruitment.