The Algorithm Doesn't Know Your CV, and That's the Problem

Why human judgement still matters when AI decides who gets through the door.
**Mulalo Tshikalange explores why human judgement still matters in an increasingly AI-driven hiring landscape**.

By Mulalo (Lali) Tshikalange| Recruitment & Search Partner and Transition Coach at Mosu Talent Solutions

Stanford researchers followed 3.4 million job seekers who submitted 4 million applications to 150 employers screened by a single AI hiring vendor. Their finding: candidates rejected by that vendor's algorithm once were significantly more likely to be rejected everywhere else it was used, regardless of their actual fit for each role. One algorithm, replicated across employers, was quietly closing doors that had nothing to do with each other.

This isn't a hypothetical. Amazon scrapped an AI recruiting tool after it learned to penalise CVs containing the word "women." HireVue's speech-assessment algorithm, used by companies including Goldman Sachs and Unilever, was found to disadvantage non-white and deaf candidates. Neither case involved a rogue vendor. Both were mainstream tools, trained on historical hiring data that simply reflected who had been hired before.

Now consider what "historical hiring data" means on this continent. It means CVs shaped by informal work, career gaps for family responsibility, qualifications from institutions the model has never seen, English that carries the rhythm of a second or third language. None of that is a flaw in the candidate. It's a mismatch between the model's training data and the shape of African working life.

We don't yet have the equivalent of the Stanford study run on African hiring data. That gap matters. It means firms here are deploying tools built on someone else's labour market, with no local research to tell them what's being filtered out or why.

Until that research exists, the discipline has to sit with people, not platforms. Someone in the hiring process needs to actually read the CV that doesn't fit the template, and ask what it's really saying.

This is the work we do at Mosu Talent. We're a human centred, pan-African talent acquisition team. No firm's hiring needs look like another's, and we don't approach them as if they do: we tailor our approach to your size, your budget, and the market you're hiring in. If you'd rather have people making these calls than platforms, get in touch.