Hiring 2.0: When AI joins the HR team

The role of AI in modern recruitment

If you’ve hired for a post – or been hired for one – in the last year, then there is a good chance that AI was involved in the process. Whether artificial intelligence is being deployed during sifting, helping write the job spec, or assessing future prospects, it has become an integral part of the HR playbook. 

But, as is the way with the introduction of many new technologies, there has been little consensus as to whether AI should be used in recruitment, how it may be deployed, or whether it is ethical to automate hiring. There are also doubts and suspicions as to its effectiveness and abilities to recognise the best candidates, as well as concerns over potential bias in AI hiring tools. 

These concerns were the subject for our 2022 review People versus machines: introducing the HIRE framework, which showed that AI hiring tools have a lot to add to the HR process, as well as some surprising strengths. Its findings also illustrate the value of HR professionals and business leaders becoming proficient in their use and understanding of the AI tools that can and will impact their businesses.

It's hard to base a study on a value judgment, so we didn’t set out to adjudicate whether the use of AI in the hiring process is good or bad, but simply whether it is better than humans, be that in hiring for performance or encouraging diversity. We wanted to see if the scepticism of candidates is justified and whether those same candidates may, to their own surprise, be better off with an algorithm at the hiring desk.

Looking at efficiency, performance, diversity, perception, how humans and AI may best work together, and the changing landscape of applying for a job, we broke down how humans and machines differ and where each excels. On the whole, we found that there was not a great deal of difference in the gains to be had by having either humans or AI in charge of an HR process, although AI definitely excelled in efficiency.

 

Where AI excels and where humans still matter

The largest shock from our research was that, although AI can be biased and prejudiced, it seems it is less so than humans and AI can actually help to iron out or remove the impact of human prejudice, making it better for diversity. As with all AI recruitment software, there are questions about the training data being free from bias, but the risks may be less than with human counterparts. After all, AI doesn’t tend to wake up in a bad mood, which could impact judgment in humans,and it won’t simply throw the final three CVs into the bin at 5pm in order to make it home in time to put the kids to bed. It can also parse 800 CVs in an hour, whereas a human may take a week or more.

That said, HR does always need a human aspect, especially when it comes to the final round of hiring. AI is some way off being able to calculate how well the new hire will get on with that awkward longstanding account manager, as well as whether they are likely to present well to external clients or members of the public. What AI hiring algorithms can do is project future performance from existing data. This is reliable overall, but human decision-making could be the factor that helps decide between two candidates who are equally matched in terms of available data. While AI may be less biased than humans, its bias is more systematic and would apply across all candidates and humans can potentially identify any high quality candidates that AI may have omitted. For the optimum HR process you need humans and AI, working together, and we aren’t close to abandoning human interviews for the final decision as yet.

While interviews appear to remain the same, it may look like the process of writing a CV has changed immeasurably – so much so that the Department of Work and Pensions is developing its own AI CV-writing tools. But pre-AI there was still a formula, whether that was something you were taught at school, or given a booklet about at university. That meant including certain things, often in a certain order and with a pre-designated layout. You were obliged to work to the formula in some way in order to get that interview.

With AI CV screening tools, the first thing that a human may get to see from your application is the covering letter, which can also be used by AI to extract and score necessary skills. But, for many recruiters, this is where you get to speak to them and show your voice and your character. It is also the place where humans work out whether a whole application has been written by AI. Despite recruiters’ use of AI, many can still be sniffy about applicants using it themselves. Although the expectation of human-only applications will wane with time.

 

The future of AI in hiring: transparency is key

This is one area where I see change coming that recognises the reality of the AI-driven recruitment process. I think it would be really helpful if recruiters were open about how they use AI when hiring for a role. This means that everyone knows what to expect, how fair the process is and how their data is used. This helps to get rid of the fear factor, as well as letting candidates know how they can use the way AI will sift to their advantage, or at least make them concentrate on key areas of importance.

Such open guides to AI usage are already being used by some businesses, with HR business HireVue publishing their ‘explainability statement’ online. As with our study, this highlights the efficiency and innate fairness of ethical AI in recruitment, as well as explaining the hiring process in detail alongside how their AI is trained.

The future of recruitment will undoubtedly include AI to an ever-growing extent, so it is reassuring to know that this is not the faceless dystopia that some may imagine. Instead, AI in HR will provide a smoother, more cost-efficient experience where HR departments can spend more time on existing staff and less on processing applications, which is priceless in large businesses and those with sizeable staff and turnovers.

 

The culture of hiring is changing, and it doesn’t have to be painful to change with it.

Find out more about LSE’s AI and Culture: Insights for Business course and read our course programme.

People versus machines: introducing the HIRE framework by Paris Will, Dario Krpan and Grace Lordan is available in full here.