Recruitment & Selection

AI and its impact on inclusive recruitment

Don’t worry. This isn’t going to be yet another long read about inclusive recruitment that tells you all about how diverse workforces are more likely to outperform their peers, and how you need to put job adverts in unusual places, or you’ll fail to attract the people you need to serve diverse customer audiences.

I know that you know that – you wouldn’t be here, reading this, if you didn’t. But knowing something doesn’t make it happen. We’re going to look at the barriers an organisation puts up against inclusive recruitment practices and how I think we’ll soon overcome them.

First, let’s look at the story so far. You’ve:

  • Got your executive leadership on board
  • Taken out all the gendered language from your advert
  • Realised that even though the Guardian is progressive, its readership is still as homogenous as the Telegraph
  • Proactively sought out ad placements in places that aren’t Guardian Jobs that genuinely reach diverse audiences
  • Actually produced a shortlist of quality candidates from diverse backgrounds that can be delivered to the hiring manager

And then, guess what? The hiring manager, who’s delighted that their unconscious bias training taught them that their favourite tactic of hiring someone as close to themselves as possible is so widely practiced it has a name, picks the person who ‘reminds me of myself 10 years ago’.

My point is this. As HR leaders we can only take the organisation so far. We must understand and appreciate that line management can and will completely ignore us in the majority of their decision making. As far as they’re concerned, their jobs and reputations are dependent on making the right hire – after all, they’re the ones who’ll have to work with the individuals. So, they’ll pick the safe option; the mini-me. It is an incredibly rare manager with the confidence and, often, arrogance to believe that they’re so good that their reputation can absorb the potential damage of a risky hire that ends in failure.

De-risking recruitment

This clearly makes increasing diversity in the workforce quite challenging. The answer is to de-risk recruitment. And this already happens. Graduate schemes are run by HR, which means that even if a line manager feels their grad is a liability, it’s not a reflection on their selection skills. The graduate can be given time, moved around and find their specialty without any risk to the manager.

And strangely, despite the concerns of bias, it’s probably Artificial Intelligence that will give us the ability to de-risk recruitment for managers. The HR and recruitment teams of the near future will be able to conduct full recruitment cycles using AI to select the best candidate.

Let’s be honest, recruitment isn’t fun. It’s also a distraction from daily deliverables, highly time-consuming, and demands that the hiring manager make an expensive, long-term decision with minimal, largely unverifiable information. In short, most line managers don’t enjoy being hiring managers, and would happily pass that responsibility off to a trusted ‘other’. Even to AI.

But, I hear you say, the AI models are biased. Well yes. And sorry to break it to you, so are every single one of your hiring managers.

And the solution to AI’s biases rests with you, and your data. Foundational models (the underlying data and code behind AI products) are being licensed or open-sourced for organisations and individuals to use and build their own technology on, using their own data.

Which means the level of bias in your AI powered recruitment journey will be, fundamentally, down to you. You will have control over the data you put in and the patterns the AI discovers. And what’s more difficult? Iterating and adjusting for the bias in your single AI recruitment model, or dealing with tens, hundreds, or even thousands of hiring managers all with completely different biases?

But that’s the destination – how do we get there?

As HR leaders it is our responsibility to identify and understand the opportunities ahead of us. AI is the biggest opportunity and the biggest change in the world of work since the dawn of the internet and will have an even greater impact.

However, having your own bespoke AI solution is still a couple of years away for large organisations and probably five years away for small and medium ones. So how do we de-risk recruitment for our hiring managers today, so that we can collect the inclusive recruitment data we need for tomorrow?

Firstly, prove the hypothesis. Do line managers in your organisation want to be responsible for hiring decisions or not? And certainly, there will be many who do, especially at more senior levels.

If the feedback is that certain levels do want recruitment responsibility, but others don’t – great. It may be a greater short-term burden, but moving activity gradually into HR now will make the eventual move to full AI recruitment much easier. But be super aware of your own biases!

Secondly, collect the data that’s important. Depending on resourcing, getting a data analyst to work on this with you can be helpful. They’ll be able to identify key metrics and indicators to be aware of that you can use to improve recruitment now and in the future.

Thirdly, look at the best things about recruitment now, and make sure to embed them into the future. It’s still a human process – people still want to engage with people, even if evaluation and selection is done by a big pile of computer code. And that’s something, if we’re honest, that we’ve not been doing well on. The AI recruiter will be able to share instant feedback on candidate strengths and weaknesses – human recruiters often end up ghosting candidates who ask for feedback. We can drag out the process far too long searching for that ‘golden egg’ and lose outstanding candidates. But we can be great at supporting candidates, building great experiences, and removing barriers.

So, think about what you’re doing well, and what you want to do better. You’ll be able to invest more time doing those things by automating the others. After all, bias in decision making is usually caused by making them in a rush.

AI isn’t the answer to bias in recruitment, but it is a tool that can help you find a solution. Easy access to AI is going to change how we work completely – you’ve probably already received cover letters and CVs written by ChatGPT. But along with threats and challenges come opportunities, if you know where to look.