Machine learning and predicting human potential in organizations

In a recent article, I showed how statistical learning – a field of artificial intelligence currently undergoing a revolution – is having an impact in a number of fields, such as law, medicine and finance. But what about the working world? Is it possible to teach a computer how to identify candidates who will be outstanding employees? Is it possible to teach a computer how to evaluate how well a skill has been mastered?

The answer is yes.

At SPB Organizational Psychology, we have used D-Teck to develop a series of assessment solutions for predicting human behaviour in companies through machine learning.

The case of D-Teck

Since 2012, we have been working on a major project to archive and structure our data. Every time a candidate is evaluated by our experts, we archive both the results of the psychometric tests and data from the consultant’s decision about the degree of proficiency in a skill set. This data allowed us to produce a series of articles about leadership in the past few years.

In 2014, we decided to take our efforts further. We wondered whether it was possible for a computer to learn how consultants make judgements about skills using psychometric data. We discovered that it was indeed possible for a computer to create predictive models and replicate the judgement of our experts in assessing potential. We decided to create a series of products that apply machine learning to organizational psychology, which resulted in D-Teck reports, available from our Boutique. We now offer a number of reports created using machine learning for assessing candidates to hire and develop.

The impact of machine learning on the job market

While this is a nascent field, its impact is already being felt. It is hard to know what the limits of these technologies will be. Will they become as intelligent as human beings? More intelligent? Will we soon be overtaken, relegated to idleness in a leisure society?

During the robotics revolution of the 1980s, many believed that robots would replace workers. But this was not the case. Factories do indeed employ fewer workers than before. Many routine tasks are performed more efficiently by robots. But factories still need employees for two reasons. Some tasks are too specialized and not routine, so they are hard for robots to do. Plus we need people to operate the machines. A number of indicators suggest that the same scenario will occur in the future with artificial intelligence and the work of experts.

We believe that cognitive tasks that are both complex and repetitive, such as making diagnoses, will eventually be done by machines. Economists Jiu and Jaimovich have already noted a drop in the number of jobs that involve repetitive cognitive tasks. This phenomenon should intensify, given that these machines are now better than people at performing precise tasks, such as image and text recognition. But, as with the rise of robots, artificial intelligence will not take away all jobs from experts. 

The limits of machines

One limit of artificial intelligence is when it comes to performing non-routine cognitive tasks. For machine learning, computers need access to a data sample. If a completely new situation arises (a situation that wasn’t in the sample that enabled the learning), the system faces an unknown situation and will not know how to process the information. The resulting predictions could be wrong. As a result, the current revolution, which is mainly based on applications of machine learning, has its limits. We can imagine that in a future revolution of artificial intelligence focused not only on learning but also on reasoning, it could be possible to build machines that reason and that can handle the unknown. But that’s not the case today.

Beyond these limits, there are certain tasks that will always require a human presence, particularly in fields related to human relations. When it comes to giving an employee or a candidate difficult feedback, supporting employees in their professional development or coordinating a team’s work, machines won’t be replacing humans any time soon.    

Articles on leadership
Four Major Contributions by Leaders to Your Organization’s Success
Competencies in Demand
5 Beliefs about Managers’ Personalities
Are Outstanding Leaders Few and Far Between?
Competencies That Set Different Management Levels Apart