If you could travel back in time and speak to a recruiter or job hunter from 10 years ago, and tell them that today's;
- CVs are written by AI.
- Cover letters are written by AI.
- Shortlisting is carried out by AI.
- Sifting interviews are carried out by AI.
They'd probably laugh at your crazy fortune-telling.
And it would be funny, if it weren't increasingly real - and if it weren't so obviously dysfunctional.
Having worked in recruitment for significantly longer than 10 years, we've seen first-hand the inexorable rise of AI in the recruitment world. Now, don't get me wrong - I can see definite benefits of AI and machine learning. However, the worry for me has been seeing too much power being handed over to recruitment algorithms – at detriment to both companies looking for talent, and for job seekers as well.
Unintended Consequences
If a critical mass of applicants use AI to create their applications, and companies use AI to select candidates, we risk creating a hiring loop where everyone optimises for algorithms instead of fit, culture, or potential.
In other words; keyword matching doesn't effectively shortlist candidates, all it tests is their ability to keyword stuff a CV.
Churn
This alone isn't great news. But the consequences could be even greater. If the recruitment industry is failing to effectively evaluate candidates for roles, will we see the already worrying percentage of people who leave jobs within 6 months increase even further?
Depending on the survey you look at, "quick quitting" statistics range from 30% – 50% of UK workers1. This has figure has been attributed to a range of issues; unrealistic expectations, a belief that the position applied for doesn't match the reality, and negative work culture. Could these issues have been addressed at an earlier stage by taking AI out of the equation?
Soft Skills
Many roles emphasise the importance of "people skills," yet automated hiring processes increasingly remove the "people" element from evaluation. Human interviewers can detect energy, humour, and cultural fit in a way algorithms still can't accurately replicate. Removing the human touch and letting AI decide whether a candidate has good people skills, risks reducing soft skills to superficial metrics that don't reflect actual behaviour.
Ironically, despite soft skills being stated as integral to the success of many roles, these characteristics usually only become apparent through context, nuance and real-world interactions - and these are the areas the recruitment algorithms struggle most with.
AI is here to stay - helpful recruitment assistant, terrible replacement
Just to reset the balance here, I'm not anti AI. And the goal isn't to reject AI in hiring, but to integrate it in ways that enhance rather human judgement rather than replacing it.
AI is incredibly efficient at streamlining logistics. High level CV sifting can very quickly weed out any applicants who don't have a specific hard skill, such as a required programming language for example.
Applicant Tracking Systems and other detection programmes can spot when AI has been used in application, flagging this to recruiters.
AI has its place, but as a helpful assistant rather than a replacement for human interaction. For those companies still wedded to their AI processes – at least be transparent about your use of AI in hiring processes. Candidates want to know when and how they're being evaluated by machines.
Why human interaction should come first
As many of you know, I will forever bang the drum that "people still hire people." Ultimately, I believe that when there is an overreliance on AI, the candidate experience suffers and companies miss out.
So many candidates I speak to are either wary of, or have had first-hand negative experience of automated processes. This is why, when we founded Matchking, we made a conscious decision to stay Human-to-Human focused, and we've never wavered from this.
The good news is the AI debate doesn't need to be an all or nothing conversation. The sweet spot sits somewhere in the middle: automation that supports better human decisions, not automation that replaces them.
