AI in Recruitment: Just Because We Can, Doesn’t Mean We Should…

We need to have a grown-up conversation about AI in recruitment.

Not a hype-fuelled panel chat with founders flogging their tools, and not another "AI will save us all" LinkedIn carousel.

A proper one.

Because I’m not anti-AI. 

I actually want more of it.

I just want it doing the right jobs. The boring ones, the admin, the repetitive nonsense that’s been part of being in Talent Acquisition for years.

Instead, what we’re seeing is the opposite. 

Shiny new tools promising “intelligent shortlisting” and “personality detection at scale.” 

What I hear is algorithms pretending they understand people better than people do. We’re handing over the most human part of hiring, judgment, connection, instinct, to something that doesn’t even know sarcasm when it hears it.

I’ve spent over two decades hiring people. You know what I’ve never seen? A spreadsheet that made a good first impression. Or a Boolean string that convinced me someone would be brilliant in a crisis. Yet that’s where we’re headed. We are treating the CV as gospel and trusting software to decide who’s worth a chat or some of our time.

And let’s be honest, most hiring hasn’t exactly been a slick, well-oiled machine up to now. Even the most structured of hiring process can break or get a little bit messy, but we make it work because there’s intuition in the gaps. We read between the lines, we notice the unspoken stuff, those awkward pauses, the unexpected jokes or the moment someone talks about a role like they actually want it.

AI misses that. Every time.

Right now, we’re seeing AI write job specs, screen CVs, filter applications, write interview questions, and in some cases, interview candidates. It’s a production line, not a process. And candidates are on to it. They know this new game and are sending ChatGPT-enhanced CVs to jobs that were likely written by ChatGPT in the first place.

It’s turned into this bizarre loop where AI is screening AI, trying to outsmart AI.

And the human being? Well, they’re somewhere in the middle, wondering if they'll ever speak to an actual person. Or whether their cover letter got shredded by a language model for not being keywordy enough. Some startups are trying to make AI “feel” more human by building agents that have little chats with jobseekers, asking you about your dream job like a robot therapist. I don’t hate it, but it is what it is, and let’s not pretend it’s replacing real connection.

Even the founders building these tools are treading carefully. They’ll tell you it’s not a total handover, more of a “smart assistant” and whilst for me that’s a step in the right direction, it feels like there is a big difference between support and substitution. And too many are nudging towards the latter.

We talk a lot about culture fit. Chemistry. Team dynamic. Gut feel.

Try training that into a model.

Tell a bot why someone who swears too much, asks brilliant questions, and completely ignored your job spec is exactly the person you want on your team. You can’t.

Tell the bot that when the CEO uses the word weasel, it’s a joke, not a literal brown rodent.

So let’s stop pretending that AI can handle the whole thing, or at least until I retire and candidates embrace AI in every aspect of our lives (why are we starting with recruirtment?). Use it to make life easier for recruiters and for candidates, but don’t kid yourself into thinking it’s a replacement for human judgment.

It isn’t.