Artificial Intelligence (AI) can radically improve employee selection processes – by enabling recruiters to assess many more applicants – but the basis of your selection decisions must be legally defensible, according to a new paper by assessment specialist cut-e.

Called Artificial Intelligence in assessment, the paper explains the role of cognitive technologies in talent acquisition. It highlights how AI is used today to enhance recruitment and how it might be used in the future. As well dispelling some of the myths about Artificial Intelligence, the paper illustrates how it can improve the speed, efficiency and effectiveness of employee selection.

“Artificial Intelligence is often perceived as a futuristic technology, but the truth is that it is already part of our lives,” said Dr Achim Preuss, Chief Technology Officer at cut-e. “The increased processing power of today’s computers, combined with the sophisticated use of algorithms and machine learning has led to an increase in AI-based applications. In recruitment, this means that huge volumes of candidate data can now be analysed and interpreted quickly and cost effectively. This paper outlines the revolutionary implications of AI. It highlights what AI really involves and what it can accomplish.”

Artificial Intelligence is currently used to automatically generate ‘test items’ in assessments and to automate scoring and reporting. Machine learning is used in data analysis to create predictive people analytics, to help employers make better talent decisions. AI’s analytical power is also used to evaluate video assessments.

“AI has been part of assessment since the advent of automated computer-based scoring in the 1980s,” said Dr Preuss. “However, AI can now further improve the precision and efficiency of assessment and it can optimise and enhance the selection experience for candidates. This means that recruiters now have the opportunity to assess many more applicants.”

The paper warns against using ‘black box’ algorithms in AI, as these can make selection decisions difficult to justify.

“The AI that’s built into talent assessment must be transparent and open to challenge,” said Dr Preuss. “With black box algorithms, it’s almost impossible to understand how they reach the conclusions they deliver. If your selection decisions cannot be explained, they can be challenged by applicants in a court of law. A better and more legally-defensible approach for assessment is to allow AI to continuously learn by ‘observing’ the best practice of human raters. When multiple human raters assess a candidate’s responses, looking for specific behaviours and competencies, an AI system can learn to understand and mimic their ratings. It will then begin to rate other candidates in exactly the same way.”

As the breadth and scope of AI expands, the paper predicts that AI will make recruitment even more efficient, productive and successful by augmenting human activity.

“AI and recruiters will work alongside each other in the future,” said Dr Preuss. “Cognitive systems perform tasks, not entire jobs. If good selection decisions are made – and candidates perform well, feel engaged and stay – then the importance of recruitment consequently rises. Recruiters will therefore become even more important in organisations as they’ll guide, coordinate, control and monitor complex AI-driven selection processes. Rather than taking resources away from a successful department, it makes more sense to supplement it with additional investment to achieve an even greater impact.”