Jobseeker claims AI discrimination behind numerous rejections from roles

Vivi Koroma Kala has lost count of the number of jobs she has applied for.

A marketing professional with 16 years of experience, she was made redundant in July and immediately started looking for a new role.

Now, two months and countless CV revisions and cover letter drafts later, Vivi is still unemployed.

‘I’ve got three kids, and the redundancy money is running out,’ she tells WTX. ‘It’s so disheartening.’

Vivi applied for a role at a large bank one Friday, just before the deadline.

By Saturday morning, she had already received a rejection email.

At another large organisation, Vivi completed a five-part hiring process over a 10-day period.

Even though she passed seamlessly from one stage to the next and was encouraged by ‘warm’ interactions with her prospective employers, she was ultimately rejected – because, she claims, of just one question she answered during the initial screening test.

‘So, one data point outweighed everything that people in the room had seen,’ she adds.

What the two experiences have in common is that their hiring processes, Vivi suspects, made use of artificial intelligence.

While she can’t know for certain, she said rejections that appear automated have certain tells – from generic phrasing (‘However, on this occasion…’) to arriving in her inbox a little too quickly.

‘It’s so obvious that nobody, no human, had assessed my skills,’ she tells us. ‘You start to see the patterns from these obviously AI-generated emails.

‘You can’t improve as a candidate when nobody actually tells you what happened. You just apply again and again, and you hit this wall.

‘It’s really dehumanising, and it does knock your confidence. Especially when you know a machine is on the other side of that decision.’

It’s no secret that companies are using artificial intelligence as part of the recruitment process.

A 2023 study by the Institute of Student Employers found that around a third of UK employers use AI in some way when hiring new staff, and that number is only going up.

‘AI is a tool, I use it,’ Vivi says. ‘I understand that employers can be under so much strain with the amount of applications that they receive, but it doesn’t mean that employers can hide behind a machine.

‘It’s their responsibility at the end of the day to be really transparent on how they use these tools, and I haven’t seen that so far.’

Vivi also suspects that, as in other applications, the use of AI in recruitment may inadvertently foster discriminatory practices in the hiring process. ‘I really can’t know, but I feel that way,’ she says.

‘Maybe not on purpose, but these algorithms can reflect discrimination, and that’s something I really, really think employers should be wary about.’

Vivi’s suspicions are supported by recent research.

A study of 2,000 UK jobseekers by the non-profit People Like Us and Censuswide found that some applicants are rejected within minutes of submitting an application, sometimes before the closing date.

The study also found applicants from non-white backgrounds appear to be rejected at a disproportionate rate.

They apply for a third more roles compared with their white counterparts but are rejected faster, at a rate of 41% within an hour of applying, versus a rate of 32% among white applicants.

Concerningly, jobseekers from non-white backgrounds are also more likely (70% vs 57%) to remove elements of their identity from their CVs in a bid to bypass bots.

AI is trained on enormous amounts of publicly available data and text.

So even if hiring managers don’t intend to discriminate, AI tools could inadvertently ‘learn’ to reinforce historic biases.

Of course, candidates can’t know for certain whether or not they have received an automated rejection; whether their application has been reviewed by a human; or whether discrimination was at play after they receive a rejection.

But researchers at UCL last year found that ethnic minority graduates from low socio-economic backgrounds are 45% less likely to be offered entry-level jobs compared to more advantaged white applicants.

This phenomenon is nothing new, and predates the AI boom. Candidates from non-white backgrounds have long reported having to edit aspects of their applications, in some cases even anglicising their names.

Sevjan Melissa Pem previously wrote for WTX about her experience of being consistently rejected for jobs for five months, having no luck across a slate of different industries.

But when she started using her more ‘socially acceptable, white-sounding middle name’ on her CV and application forms, the interviews started rolling in.

‘I scheduled four interviews in one day; one company phoned to set it up the same day I applied. It was almost too easy.

‘There was only one explanation for what happened with my job hunt: outdated, accepted, inherent racism.’

Comedian Jenan Younis also told us how she got more comedy bookings when she adopted a more ‘white-sounding’ pseudonym.

‘Over the course of [a year] I emailed comedy promoters across the UK alternating between “Jenan Younis” and “Janine Young”, with the content of my comedy CV and clips remaining identical,’ she wrote.

‘Several hundred emails later, the results spoke for themselves. “Janine” was overwhelmingly more successful in securing a gig offer, with 81% of promoters replying, while just over 12% responded to my emails as “Jenan”.’

Vivi can’t wait to find the perfect job any longer and has instead launched her own consultancy project while freelancing to supplement her earnings.

She also wants more applicants to know they can ask companies to give them human feedback if they receive what they believe is an AI rejection.

‘I didn’t know that I could ask for a human to have a look at my application,’ she tells WTX.

‘And I don’t think I’m unusual in that. I’ve spoken to many people, and a lot of people are facing the same thing.’

People Like Us has now launched a new campaign to expose what it calls ‘automated discrimination’ in UK hiring.

Reject the Rejections is a national campaign launched on Wednesday, September 16, aimed at making hiring more equitable.

It has also created a free email template developed with Planes Studio that rejected candidates can use to ask whether an automation was used on their application and to ask for a human review.

You can find out more here.