Continuing his series exploring AI and its impact on the new world of work, Morgan Philips Founder & CEO, Charles-Henri Dumon considers the benefits but also the limitations of these scientific techniques on recruitment and talent attraction.

“Machine learning, deep learning, robotics, algorithms, chatbots, structured and unstructured data, supervised and unsupervised learning – it can be easy to get lost in the quagmire that is artificial intelligence terminology!  

Sourcing capability

“Definitions aside, we’ve known for some time that AI can make the life of the recruiter a lot easier. Rather than a person spending hours on end sifting through CVs, we can get a machine to do it, using algorithms to match people to jobs. Programmed to pick out relevant keywords that you’re looking for in our specific roles, CVs can be quickly and efficiently sorted to produce an effective, suitable shortlist.  

As far sourcing and identification go, AI can make a big difference in cutting through the huge volumes of data, given the thousands of CVs and applications that some global companies receive. Predictive analytics technology, based on historic data of previous successful hires, allows recruiters to make even better hiring decisions. The time saving is significant and means that the recruiter can focus on other key parts of his or her role, for example meeting clients fact to face and relationship building. 

Eliminating bias

Then there is the question of bias, which has always been a bone of contention the recruitment process. Whether age or name, there are many different ways in which a candidate could be discriminated against, both consciously or unconsciously. There have been well documented cases of candidates changing their names on their CVs and being treated differently as a result. So-called ‘blind’ recruitment is now being encouraged as a result. 

But does AI fare any better? Although it has the potential to reduce bias, it’s only as effective as the information that it’s fed. So in actual fact, depending on the data used, algorithms can actually magnify that bias, therefore defeating the purpose of the exercise. Amazon infamously had to pull the plug on its AI recruiting tool back in 2018 because of gender bias – women were being excluded for its software developer and other tech jobs. 

Human intelligence

We now have facial expression software that is being used to interview candidates by analysing aspects of their behavior such as body language, word choice and speech patterns to assess against requirements deemed to determine if a person will be successful in their new role. Machines though can only draw conclusions on the basis of past data provided and so you’re inevitably going to recruit the same or similar type of person for the pre-defined traits that have been established. This doesn’t bode well for diversity, creativity and innovation.  

As the French philosopher Gaspard Koenig states in his book, ‘La fin de l’individu – un voyage d’un philosophe au pays de l’intelligence artificielle’, machines can’t conclude that a candidate ‘ticks all the boxes’ in the same way an interviewer would. As much as AI is the future and does have a part to play in recruitment, we mustn’t lose sight of its limitations. 

Even with more controls in place, we still haven’t reached a point where it can substitute a human’s level of thinking. At least not yet!

Read part 2 – the impact of AI on tech

Learn about the three waves of AI