Artificial intelligence (AI) is making headway into a variety of industries and is, according to Gartner, on course to become the next IT industry megatrend.
One industry where the use of the technology is being actively explored is recruitment, where enterprises are drawing on its capabilities in various ways to help them find new staff.
According to Martin Ewings, director of specialist markets at recruitment consultancy Experis, the technology is in common use, but industry-wide adoption is still some years off yet.
“We see technology already in the mainstream, allowing recruiters to increase the value add that they have into the process, while ensuring anyone who doesn’t make the selection feels like they’ve been assessed and engaged with in a correct, professional way,” he says.
“You’re looking at three to five years before it’s totally adopted because there are still organisations reluctant to remove the human interaction side of things,” he adds.
Achieve efficiency with AI
Olly Burns, product director at online recruitment firm Totaljobs Group, says one area where AI comes into its own is the efficiency it brings to sifting through large numbers of applications.
“Where it really has value is in helping the recruiter shortlist the candidates,” he says. “They want to get to the shortlist of the most appropriate candidates in the quickest time possible. If we can help them do that faster [using AI], then that’s a win for us.”
Andy Heyes, managing director of technology-focused recruitment firm Harvey Nash PLC, says AI could also speed up the interview booking and coordination part of the job-seeking process.
“There’s no reason why that can’t be automated,” he says. “That [process] is prone to human error, so there will already be products out there, [such as] virtual assistant systems that can provide that service.”
Personalising the job search
Some organisations are using AI in the form of chatbots to personalise and improve the job-seeking experience for candidates too, with the software able to offer candidates a specific, conversational response to any job application they submit.
Online job site Monster is one of them, and it plans to introduce a chatbot to its recruitment process over the next couple of months that will use rule-based grammar and machine learning to personalise the application experience for jobseekers.
Sinead Bunting, European director of consumer marketing at Monster, says the aim of the technology is to help candidates on their journey when applying for jobs.
“It’s more of a friendly and engaging experience,” she says. “If we can offer that to our users and they feel as if they’re getting that personal experience and they’re getting the information they need, then fantastic.”
AI can also play a role in helping manage candidates’ expectations about how long they will have to wait for a response to their application, says Burns.
“If an AI system has access to a recruiter’s communications, it can learn, for example, how long it normally takes a recruiter to contact a shortlisted candidate. It can therefore set the candidate’s expectations and inform unsuccessful candidates automatically at an appropriate time.”
Receiving slow (or even no) response to an application is one of the biggest complaints recruitment firms hear about the services they provide, Burns continues.
“[That’s] especially [true] at the beginning of their journey when they’re interested in a job or they’ve just applied for a job,” he says.
“Artificial intelligence and that question and answer paradigm could help there, and you could start using that kind of technology to provide feedback to the candidate.”
Read more about artificial intelligence
- Pitting a chess computer against a chess champion used to be the end goal for machine intelligence. We explore how game play has improved artificial intelligence.
- Revelations that researchers at Facebook had to switch off two bots that went rogue have raised questions about the safety of artificial intelligence.
Another complaint often levelled at the industry is that recruiters are too reliant on grades and qualifications to decide who the best candidates are, says Jeremy Hindle, CTO and co-founder of machine learning-based recruitment app Headstart.
“It’s very difficult to measure how capable somebody is, given the existing structure of a CV and the way recruitment systems within businesses currently work,” he says.
“They are very heavily weighted towards grades and where they’ve studied, and I don’t think that’s the best way. Companies are missing out on some of the best talent. It’s most apparent in sectors like technology, but also in advertising and media,” he adds.
It is an issue Hindle and his Headstart co-founder, Nicholas Shekerdemian, are seeking to address with the app, which uses machine learning to find relevant roles for jobseekers.
The candidate starts by creating a profile on the website, including details of their personality traits. This additional information is not shared with the hiring companies, but allows the Headstart algorithm to find suitable roles for the candidate, based on a wider range of factors.
Assessing the short-comings of AI
While there are benefits to incorporating AI into the recruitment process, there are drawbacks too, particularly when it comes to selling a job to candidates.
“A hiring decision is a two-way process,” says Kevin Green, CEO of industry trade body the Recruitment and Employment Confederation (REC).
“The candidates want to make a decision about whether this is the right role for them, whether they like their potential boss, whether the people in the team are people they can work with and that they think it’s the right environment, as well as the right set of tasks that they would be asked to do.”
In situations where candidates with certain skills are in high demand, and could effectively pick and choose who they want to work for, it often comes down to the recruiter’s ability to sell the opportunity to them, which is probably beyond what AI can do, says Ewings.
“Particularly with rare skills like IT security, big data roles, some cloud-based and digital roles, these candidates are in demand. Why would they choose your organisation above another? Yes, it’s great that an organisation can use AI to sift candidates out, but the candidate has to have a warm experience as well and has to be able to judge the organisation,” he adds.
Burns thinks AI technologies will never completely eradicate the need for human intervention during the recruitment process. “Deciding who to hire is a big deal as well, it matters hugely to employers and quite rightly so. To do that effectively, you need human-to-human interaction,” he says.
This is a view shared by Harvey Nash’s Heyes, who makes the point that recruitment is a “people-centric industry” so it stands to reason that humans will always have an important role in the process of hiring.
“[AI] is coming but there will always be a human element and there has to be because it’s dealing with people. People are unpredictable and people change roles and look for different things throughout the process,” he adds.
Tackling unconscious bias through technology
Introducing AI to the recruitment process has the potential to help tackle the problem of unconscious bias blighting the decisions human hiring managers make, but that all comes down to who is responsible for training them and the data they are fed.
Kevin Green, CEO of REC, says the person influencing the technology must be chosen carefully to avoid this happening in tests. “Or what you end up with is that you just hardwire some of the prejudices of the individual into the assessment technique,” he says.
Headstart CTO Jeremy Hindle makes a similar point, adding that – if the technology recognises the company has been employing a particular type of candidate time and again – it will continue to specifically look for that kind of candidate in the future
“They can essentially recreate a recruiter. However, that’s really dangerous because you train biases into that algorithm, you train all of the problems you’re associated with in a particular recruitment process and it becomes even better at them in a way. This means it gets worse. That’s what bad machine learning would look like in recruitment,” he says.