Artificial intelligence has become a bit of a buzzword of late. It’s been hailed as transformative technology that is set to revolutionise society, but people are still sceptical. With mundane, repetitive tasks in the workforce already being automated, humans are being displaced and their skillsets made redundant – it seems that artificial intelligence gives with one hand and takes with the other.
Automation isn’t bound to the constraints of those monotonous musings, however. As AI develops in complexity, with implementations such as machine learning and deep learning, it starts being able to make decisions for itself and begins acting in a more human way – however dystopian that may sound. In short, this means that jobs and processes that tend to require a softer approach and a connection between people, hiring and executive search for example, are not out of reach for AI systems.
It could be in 20 years’ time, it could be more or less, but it is almost a certainty that we will reach a point of Technological Singularity, completely automating the use of AI without the need for human input. Machines currently rely on the input of datasets, but as they begin procuring more information themselves based on that initial data, they will eventually be able to make rational decisions using their own intelligence.
This eventuality will reshape any given industry entirely, recruitment included. It will mean that hires can be made without the candidate ever coming into contact with another human being until their first day on the job. It sounds obvious that this is by no means the best approach, but it doesn’t seem to have been considered in the conversation of automation in recruitment, an industry which does not have the best reputation for putting candidates first. Thankfully, there are many within the industry who are challenging that perception but it’s important to remember to put people first in the digital age. Candidates want to feel valued, and if the prospective employers can’t spare the time to meet them then it comes across as a sign of things to come.
AI in recruitment is not entirely new. Some employers have been supplementing their hiring process with the technology for a couple of years now, most commonly to screen candidates before the interview stage. Used in the right place at the right time, it fulfils a key purpose and makes the process more efficient. However, when one of the world’s largest employers tried to create a platform that analysed existing employers to then find those characteristics replicated among its applicants, the 100,000 staff datasets they had recorded proved insufficient to build a successful model. Gargantuan pools of data are needed to build an accurate platform, but hiring practises have been flawed and biased in years gone by, so it isn’t possible to backdate the information – that’s where Amazon got caught out.
In many ways, the Singularity is already beginning to take place. Currently, the biggest stumbling block for people-led development of AI is that we cannot always understand why it has made certain decisions; machines demonstrate no empathy and no sign of ethical reasoning, so we need to know how they are coming to their conclusions in order to tweak algorithms and ensure the right outcomes are being reached. This is an issue that stems from the disconnect between data scientists and those working in the recruitment industry – there is almost no overlap between the two. Recruiters are aware of the changes that need to be made but can’t carry them out, and vice versa for data scientists. Without this collaboration, we’ll get nowhere.
About James Wright
James Wright is a Senior Associate at Carmichael Fisher. At this month’s WEI International Academic Conference at Harvard Faculty Club, Boston, USA, James will be presenting his research project on the introduction and influence of artificial intelligence within the recruitment industry and specifically establishing the impact on employers and candidates, with the view of publication.