The Future of AI and Diversity in Recruitment
Artificial Intelligence (AI) is too often portrayed in the media as a damaging force that is set to wipe out many jobs. Yet the fact is that it could also prove to be a force for good in terms of improving diversity, inclusivity and employee engagement.
A number of AI start-ups are developing ways of using the tech – which is basically a series of self-learning algorithms – to eliminate the thorny problem of “unconscious bias” in the hiring process.
One thing is for sure, AI is certainly going to reshape the role of the recruiter in the so-called fourth industrial revolution, with one recent survey claiming that 87 per cent of recruiters are excited about working with AI in future.
HR, Diversity and the Globotics Upheaval
Economist Richard Baldwin predicts that the forthcoming “globotics upheaval” is going to disrupt professional white-collar jobs in the same way that automation and global trade disrupted western manufacturing in the late 20th century.
Yet Baldwin’s vision is in no way that clichéd dystopian future where we all lose our jobs to robots, as he also persuasively argues that future jobs will be far more human, involving much more social interaction and face-to-face contact, as we transition to a world in which software robots and tele-migrants do everything else.
Which is exactly why workplace culture, diversity and employee diversity and engagement are all becoming far more important for recruiters and HR departments alike.
Companies are people
“Companies are nothing without the right people, but leaders are beginning to realise that a lot of their human resources and human capital practises are outdated,” says data scientist and AI specialist, Mike Bugembe.
Bugembe praises LinkedIn’s recently launched Representative Results AI feature as a great example of how AI can add value to all parts of the business.
“The processes used to find, recruit and retain talent, are fraught with human cognitive bias – where a person’s background, personal experiences, social stereotypes and cultural context will impact their decisions and actions without them realising it.
“For example, it should be no surprise when recruiter who attended a red brick university or who went to a private boarding school, is inevitably drawn to CV of someone with a similar background.”
Unconscious bias underpins discrimination
This type of unconscious bias underpins traditional methods that continue to have a negative impact on diversity. Which is exactly where the fascinating HR tech start-up pymetrics sets out its stall.
Pymetrics found that the traditional hiring process is “broken” as it leads to women and minorities being at a 50 to 67 per cent disadvantage.
And following the news late last year that Amazon ditched its own AI CV-reviewing software, because of the fact that it was reinforcing discrimination in the hiring process, pymetrics’ EMEA lead Tom Viggers wrote recently on the ways in which, “the use of algorithms to select candidates, or indeed any kind of pre-defined criteria, has its limitations in the world of recruitment.”
AI may well offer efficiency, but when it’s based on historical data and traditional hiring processes it is soon shown to replicate past inequalities in HR and recruitment.
All of which led the pymetrics’ AI expert to pose the ten-billion-dollar question: “can there ever be a definitive way to review applicants efficiently and fairly?”
AI unearths candidates with true potential
One person who is adamant that AI is able to eliminate the thorny problem of unconscious bias in the hiring process is Mojo Mortgages, CEO Richard Hayes, who argues that, in big business with 40,000-plus employees: “AI can be used to support the recruitment process by analysing the characteristics of their employees such as gender and ethnicity. Subsequently, AI could be then be used to fill in the gaps, thereby creating a more diverse and fairer workplace.”
Hayes is also very certain that diversification and policies to support a fairer landscape “should absolutely be the responsibility of businesses leaders and, as with most use cases, AI is simply there to support a more intelligent and efficient process.
“Utilising algorithms to determine what characteristics are required to succeed in a certain role, has allowed companies such as ourselves to massively evolve our recruitment strategy, enabling us to make far more informed decision about a candidate over and above our gut feel of whether they’re the right candidate or not.”
Hayes cites Workable people search as one specific way of utilising AI to unearth great candidates, noting that, “without the use of this AI led service, we maybe would never have found certain outstanding people.
“AI helps to unearth the true potential in a candidate and eliminates the risk of hiring based upon relationship, or prior experience.”
GIGO: bias is in the data, not the AI
Algorithms of course are only as useful as the data you feed into them. Which means there is always the challenge of “garbage in, garbage out”, or GIGO, as the old computing adage has it.
To put this in terms of recruitment and HR, if the input data on candidates for a job is biased, then the outcome will also be biased. So, if the past historical data is heavily skewed towards college-educated young white males, then the AI will continue to recommend those same candidates.
As Mike Bugembe explains further, referencing Amazon’s AI PR fail last year: “Whilst AI may not have any biases in the way it makes a predictions, prescriptions or helps a decision, it can reflect the biases that exist in society simply because of the fact that it learns from the available data.
“The constraint therefore is not on the way the machines or algorithms predict, it is on the data that is used for the machine learning process. This is common knowledge for any data scientist and with all of the expertise at Amazon, they really should have known this before they built the algorithm. Never the less, it should serve as a cautionary tale for any organisation planning on using AI – companies need to ensure that it doesn’t reflect todays imperfections and prejudices.
“I still believe that we should absolutely be looking to make a case for a more AI driven HR because companies are nothing without the right people. Companies need to put intelligent systems and processes to find, recruit and retain. Therefore, I would look at the machine learning bias in another way. Whilst we can and we must begin to collect a wider set of data, we can also change the variable that we use to tell the machine that it has been successful.
“Currently most algorithms are trained to find the right skills irrespective of a person’s gender, age, disability, ethnicity, sexual orientation, gender identity or social background. This means that the algorithm will assume that the company will benefit from having employees that are the same. However, since diversity brings new perspectives, the algorithms should be trained to find people who will bring a new perspective but will still have the minimum required skills. Shifting the focus from skills exclusively to asking the machine to predict perspective, (the real benefit of diversity).
“Collect data on candidates that are likely to give a new perspective and thus produce better solutions to the workplace then train the machine to find candidates that are likely to give a different perspective.”
Pair AI with human expertise for transparency
Adrian Ezra is the founder and CEO of HR-tech start-up JamieAi, a UK-based HR-tech start-up which has set itself the lofty mission to be the most accurate job matching platform in the world, “pairing AI insight with human expertise to deliver transparent and unbiased hiring for data professionals, with the aim of fixing the broken recruitment sector.”
With AI expected to create $13 trillion in value for businesses by 2030, Ezra identified what he sees to be a clear need to create a universally integrated job matching tool powered by AI to save businesses millions in recruitment.
“While other industries are rapidly adopting AI and machine learning, recruitment is clinging rigidly to the old, profligate ways of working that benefit neither job seekers or businesses,” says Ezra. “Our vision for JamieAi is to dramatically increase the accuracy of job search and in doing so, save businesses huge amounts of time and expenditure. A client can save more than £225,000 and over 600 man-hours, when compared to traditional methods of recruiting.”
With the UK recruitment sector currently worth £35billion, it’s clear why there is such an appetite for deploying these kinds of AI solutions in HR.
Social interaction is going to remain the one area where humans will trump robots for many years to come. Which why eliminating our own unconscious biases in our hiring practices is vital.
Workplace culture, diversity and inclusion is growing in importance, which is why there is a healthy number of exciting new tech start-ups (including all of those mentioned above) that are developing ways of using AI to boost workplace diversity and inclusion, in addition to streamlining the hiring process to eliminate inefficiencies.