AI and recruitment: Are we in a race to the bottom?

It’s my first job interview in more than eight years. Even though it’s a video interview, I’m still keen to impress.

When I log on, my interviewer, whose name I didn’t catch, looks relaxed and friendly.

He asks carefully articulated questions, listens intently, and even asks follow ups regarding particular examples I mention.

But then, strange things start happening.

He takes a while to process what I’m saying, and his facial expression remains unchanged. Then, halfway through asking me to explain a particular work scenario, he disappears without another word. He’s an AI – and he’s crashed.

If this is the future of job interviews, it’s off to a shaky start.

When it works, the point of using AI video interviews is to relieve the burden on companies’ HR teams, who are being inundated with applications since the UK jobs market hit an all-time low.

The number of vacancies available are down 12% compared with this time last year, according to the Office for National Statistics, causing the number of applications per role to rise by 65%, according to data from recruitment platform Tribepad.

The company behind the technology I used, recruitment platform Test Gorilla, says the glitch I experienced hadn’t happened prior to that day, and impacted a “very small number of candidates”.

Candidates in a real process would be able to log their issue with the on-screen help widget and receive a new link to restart the assessment, the company’s VP of product, Claudia Baijens, says.

Test Gorilla is already working with recruiters like Talent Solutions Group to screen candidates using AI video interviews.

The AI video interviews give candidates a score so hiring managers can work out who they will actually see themselves.

“It allows us to be able to prioritise and speak to that 10% of people that we actually want to reach,” says Natalie Jafaar, a principal consultant with Talent Solutions Group, who is based in Sydney, Australia.

Video interviews conducted by AI are not yet commonplace. But other AI tools are more common in the recruitment process, for example, in writing job ads, filtering CVs, running skills assessments, sending email responses and scheduling interviews.

Homecare provider Cera launched its video recruitment tool Ami in August.

Rather than video, it uses AI to conduct interviews with candidates over the phone.

It estimates that Ami saves two days a week of human recruiters’ time in processing the roughly 500,000 job applications it currently receives each year to join its 10,000-strong workforce.

Ami has now helped it recruit over 1,000 people and Cera says it has reduced the company’s recruitment screening costs by two thirds.

But what do jobseekers actually make of interacting with AI rather than people?

Jim Herrington applied for more than 900 hundred jobs after being made redundant last year from his role as a marketing director for an electronics company.

He argues that AI screening looks for certain keywords in an applicant’s CV. As a result, the bigger picture, which might reveal whether the applicant is actually a good fit is lost.

“Recruitment agencies aren’t necessarily doing their job properly, because they’re just using software and they’re not actually physically looking at applications,” says Suffolk-based Mr Herrington.

He has since secured a role as marketing and communications director for health testing company Omega Diagnostics.

He also argues that AI video interviews could be bad for the reputation of the companies that uses them.

“If a business hasn’t got the time or courtesy to speak to me themselves, then I’m just not interested,” says Mr Herrington.

“In an interview, there would be so much that an AI just cannot experience. For me, it shows a total lack of respect to the candidate who has spent time and energy in applying.

“Where does it stop? We need to value our employees of the future – not to subject them to experiences like this.”

Job hunters should also beware that scammers are using AI to promote and conduct interviews for fake jobs – asking for money for bogus training or equipment.

Mr Herrington says he has experienced this himself, having received phone calls from robotic voices. He hangs up immediately.

Mr Herrington resisted using AI to craft his CV, covering letters and presentations.

But AI is enabling job seekers to apply for more jobs than ever, even ones they aren’t qualified for.

“There are bots that can apply to 1,000 jobs on your behalf while you sleep, tailoring your CV for each of them. This means that the fewer jobs that are on the market are getting more applicants than ever, not just because there are more candidates, but because these candidates are applying to more roles than ever,” observes Lydia Miller, co-founder of Ivee, a recruitment platform focused on career break returners.

Ms Miller is concerned that the proliferation of AI on both sides of the recruitment process is causing “a race to the bottom”.

“This just creates a really dire state where the only way that recruiters and companies can possibly sift through these is by using AI to filter them, so a lot of people are just getting automatically rejected or ghosted from roles. That is less to do with their actual skills, because no human has seen their CV,” says London-based Ms Miller.

Ms Miller also warns that candidates will end up preparing themselves to say what the AI wants to hear, rather than communicate their actual skills and qualities.

“It’s almost the same as with exams at school – you learn to answer the question to get the marks. In the same way people are keyword stuffing their CVs to get past AI screeners, they’ll be keyword stuffing those AI screening calls. There’ll be people on TikTok talking about how to hack your way through a first round AI interview,” Ms Miller predicts.

Ms Miller acknowledges that AI video interviews can be beneficial for people who are neurodivergent, or simply introverted, and might feel more at ease talking to an AI than a human.

But there is also the problem of potential biases that an AI platform may have inherited from how its programmers have developed it, in terms of the data it has been fed and the rules it has been set to operate by.

“Companies are filtering CVs with imperfect AI that has bias inherent in it,” says Ms Miller.

It’s an issue for Annemie Ress, a former chief HR officer who now runs her own talent development consultancy, PurpleBeach.

Ms Ress says AI video interviews would not be able to judge a candidate based on everything the employer is looking for, only the criteria it has been programmed with.

“A seasoned recruiter or hiring manager will always know within the first interview what to explore or not. I think AI works if you have good checks and balances throughout the process, and is treated as one perspective,” she says.

“But I think it’s too early to place all your reliance on AI. It’s a balance between efficiency – and potentially losing out on great talent.”