You’re hired, said the machine
If we introduce AI at the start of the hiring process, will we also do the same at the exit stage?
Processes which require intense human effort are often difficult to scale. Think about hiring. Many organisations are faced with high churn rates due to employee dissatisfaction with the role, or perhaps lack of organisational-employee fit. Either way, recruitment is constantly on the mind of every manager. Every recruiter will tell you that perhaps the hardest part of their job is shortlisting suitable candidates.
There have been countless times in my career when I asked for a list of suitable candidate resumes after the initial paper-sift, and lo-and-behold, not a single person was appropriate for the role. It’s not that they were all necessarily poor-quality candidates, it’s just that they weren’t matched to the role I was hiring for. This leaves the candidate in limbo, the hiring manager frustrated and the recruiter asking, “where did I go wrong?”. Is it time to automate the hiring process entirely and hand it over to a machine learning algorithm or artificial intelligence (AI) engine?
The HR team at Unilever thought so. They wanted to diversify the workforce and widen the pool of those who applied for entry-level roles. To achieve this, they implemented a new recruitment process so that in the first round of interviews, candidates were asked to play an online neuroscience game which assessed traits such as risk aversion.
If they passed this phase, then in the second round, a video recording was made while the candidate answered more specific role-based questions. The AI examined the recording, assessing content, intonation and body language. The best candidates were invited to attend a third stage interview at Unilever offices with humans who made the final decision.
As prospective candidates could easily access the system, applications in the US alone soared from 15,000 in the previous year to 30,000, with a broadening of socio-economic diversity. In addition, the average hiring time went down from four months to four weeks and recruiters saved 75 per cent of the time from the hiring process.
Unilever has said that it has saved 100,000 hours of recruitment time in the last year for their graduate hires, resulting in over $1m in savings. Other multinationals have used similar tools for entry level and graduate roles.
As a hiring manager, I can certainly see some of the operational benefits, however I’d also have some concerns about the whole process: as the algorithms which run these programmes are designed by a limited – and quite often – homogenous group of people, there could be biases built into the programming of the AI. In addition, hiring by AI doesn’t come across as a very opaque process since it misses the soft human skills required to convey to a candidate why they weren’t selected for a role.
But most of all, what concerns me is that if we’ve introduced AI at the start of the hiring process, will we also do the same at the exit stage? Will an AI tool be asked to fire someone as well? Will there be a version of an earlier incarnation of Donald Trump, the one on The Apprentice, who told would-be employees “You’re fired”.
If that happens, how big a leap of the imagination would it be to find an AI unit sitting in the West Wing, doing a lot more than hiring and firing folk? Let’s hope that science fiction remains just that and, in the meantime, I receive a suitable list of candidates from my HR team.
for almost 40 per cent of all e-commerce sales in the US last year. This shift of marketshare and the related demise of physical retail should continue and might even accelerate in the future. Companies that are operating in physical retail without a thought-through complementary online sales strategy are going to face even more headwinds going forward.
While overall retail sales are in the trillions, global e-commerce penetration was only around 11 per cent last year, according to Euromonitor. This indicates increase in retail stores closures in the US last year, versus 2018