It’s time for more radical innovation in recruitment. Hiring practices have barely changed in the past 50 years and are not only cumbersome and inaccurate but have also not kept up with the changing nature of work.
So say Tomas Chamorro-Premuzic, professor of business psychology at University College London and Columbia University; Franziska Leutner, lecturer in occupational psychology at Goldsmiths College, University of London; and Reece Akhtar, co-founder of Deeper Signals – all experts in psychological profiling and talent management. And, they believe, now is the time for that radical innovation. For, as employers move from using digitalisation to automate tedious processes towards using artificial intelligence (AI) and machine learning to enhance or make knowledge-based decisions, the potential for improving recruitment and selection is huge.
And improving recruitment is vital if employers are to unlock and harness human potential, for today we live in a world where most people are misunderstood – by both their employers and themselves. The good news is that there’s a robust body of psychological knowledge and science to significantly increase the probability that every individuals ends in in the right career, maximising their fit between their unique disposition, interests and talents and the particular requirements of the job or career they pick.
“We need to bring the science of recruitment to life. The better you know yourself and the better organisations know you, the more likely you will end up in the right career,” say Chamorro-Premuzic, Leutner and Akhtar in their new book The Future of Recruitment: Using the New Science of Talent Analytics to Get Your Hiring Right.
“Anywhere in the world in-demand workers and sought-after talent will prefer to join organisations in which they are appreciated for what they contribute, understood and valued for whom they are, where politics and nepotism play a marginal role compared to meritocracy and where the culture rewards moral and prosocial behaviour while sanctioning toxic acts,” they say.
Humans are terrible judges of talent
The problem is that, despite organisations’ recognition of talent as a key differentiator for success, today’s talent practices are built around outdated, unscientific and ineffective tools. And these tools are generally far from accurate, useful or fair and, in many instances, biased and corrupt.
For example, humans are terrible judges of talent, say the authors. We regularly impose our own biases into the job selection process, even when we try not to. We often value intuition over data. And the job application process itself is dehumanised.
“Companies often give away very little information until the last steps of the recruitment process, especially about the team and people an applicant would end up working with. Compare this to the online dating world and imagine you would need to submit your life history and undergo automated tests before even finding out who you might end up on a date with,” say the authors.
In contrast AI algorithms can help recruitment managers enhance their knowledge and expertise through providing objective insights to counter their bias and limitations. Technology that captures data will provide these science-based insights, enabling recruiters to take evidence-based decisions to gain competitive advantage in the war for talent.
Among the technologies the authors identify are:
• Video interviews
• Game-based assessments
• Social media and web browsing data.
Data mined from tools like these can improve current assessment practice by building highly accurate profiles to shape hiring decisions in seconds, say Chamorro-Premuzic, Leutner and Akhtar, as every single one of us produces thousands of ‘talent signals’ every day – each revealing something unique about our personality, talent and experiences.
Take social media. It is estimated that 84% of firms are using social media sites for recruitment, 44% are using social media profiles to screen candidates and 36% have disqualified candidates on the basis of the information they have found (Society of Human Resources Management). However, these human decisions are subject to the same ‘in’ and ‘out’ group biases that beset interviews and other judgements of talent.
AI and HR can work together for better outcomes
AI, on the other hand, can surpass human evaluation:
• Digital platforms and devices can objectively measure behaviour, removing the need for biased human evaluations
• AI algorithms can be optimised to maximise the prediction between our digital records and performance indicators
• AI’s ability to scale and reach a greater number of candidates from different backgrounds can create more diverse and efficient hiring practices
• AI algorithms can provide greater transparency into how data are used and weighted to reach a hiring decision, thereby producing fairer and more ethical practice.
It is therefore easy to see that the future of HR will become “algorithmic,” say the authors. “Automated workflows will take care of the traditional and everyday tasks while data and AI will become central tools in talent decision making.”
But they warn: “HR leaders need to begin the work of upskilling their teams otherwise they run the risk of understaffing their organisation and losing talent to competitors.”
Three competencies HR practitioners must develop
To ensure emerging technologies are implemented for the best outcomes, HR professionals need to become more comfortable with three competencies, the authors believe:
1. Tracking performance and business impact
Unfortunately, say Chamorro-Premuzic, Leutner and Akhtar, enthusiasm for predictive new technologies does not always go hand-in-hand with the implementation of basic performance management practices. Often recruiters do not know how to define talent or performance, nor have they operationalised these definitions into objective and high-quality measures of performance. Without outcome data to point their predictors at, data scientists will not be able to train reliable and valid talent algorithms. HR leaders need to spend the time investing in improving their organisation’s ability to collect accurate performance metrics and more closely attribute job output to individuals.
2. Data literacy
HR practitioners will need to become more data literate if they are to turn data into the insights and knowledge needed to improve decision-making and communicate business requirements to the data scientists. In other words, they need to develop the ability to interpret data and reach valid conclusions using statistical techniques. Otherwise, say the authors, HR could become the bottleneck to the use and availability of rich talent signals.
3. Algorithmic responsibility
HR professionals need to become mindful of how algorithms are used and deployed. This means developing skills that allow them to select the right providers and vendors, critically evaluate the effectiveness and limitations of assessment algorithms and know when safeguards need to be put in place to detect and remove bias. As algorithms become further intertwined in hiring practices the authors expect HR departments to employ a head of HR ethics to safeguard against the risks.
HR leaders are in a unique position to harness new technologies to transform the way they select talent. The key challenge, say the authors, is to start building the in-house expertise that is needed to use the tools to their full potential and ensure AI-based recruitment technologies do not turn into what American data scientist Cathy O’Neil calls “weapons of math destruction”.