The Predictive Power and Cognitive Limitations of AI in HRby@liudmyla.semyvolos
218 reads

The Predictive Power and Cognitive Limitations of AI in HR

by Liudmyla SemyvolosMay 31st, 2019
Read on Terminal Reader
Read this story w/o Javascript
tldt arrow

Too Long; Didn't Read

AI and HR seem to be a match made in heaven — in some innovative heaven, driven by smart technologies. Intelligent apps and chatbots have penetrated deep into the processes associated with managing human resources. They are here to meekly do all boring repetitive tasks like screening resumes, answering common questions or onboarding new employees.

Company Mentioned

Mention Thumbnail
featured image - The Predictive Power and Cognitive Limitations of AI in HR
Liudmyla Semyvolos HackerNoon profile picture

AI and HR seem to be a match made in heaven — in some innovative heaven, driven by smart technologies. Intelligent apps and chatbots have penetrated deep into the processes associated with managing human resources. They are here to meekly do all boring repetitive tasks like screening resumes, answering common questions or onboarding new employees.

Can we expect more from digital hard workers? Of course. The new generation of AI technologies is supposed to ensure a personalized approach to each employee, forecast the workforce behavior, and help HR managers with detecting, motivating, and retaining talent.

Will a particular candidate be a good match for the position? What should be done to increase performance across the company? Who is going to leave your company in the near future? Who is ready to become a team leader? Smart digital assistants can potentially generate intelligent data-driven answers to these and other complex questions. In this article, we’ll review the most advanced use cases of artificial intelligence in human resources management and discuss some of AI’s limitations.

Hiring the right employees

AI has already gone far in hiring people. The achievements of conversational chatbots like Mya are really striking: they can answer up to 75% of the questions from candidates while 73% of interviewed people do not even realize that they are interacting with a bot, not with a human agent.

When it comes to choosing candidates with the right qualifications, technologies have also proved to be extra helpful. The point is, our decisions are often driven by emotions, and not always for the better. As one survey revealed, one in five UK bosses decides whether or not to hire an applicant within one minute of a job interview. Sixty seconds are hardly enough time to objectively evaluate a person’s skills and qualities. Employers judge prospective employees on first impressions, based on appearance, clothing, tone of voice or even a type of a handshake. None of these parameters are relevant to education, problem-solving ability or professional experience (with a few exceptions such as hiring models for the beauty industry).

AI algorithms are not affected by emotions, so they can painstakingly sift through dozens or even hundreds of applicants to detect qualified candidates. For instance, the Hiredscore recruiter’s AI assistant helps to find the right person faster and reduces cost per hire by leveraging big data and predictive analytics. AI-fueled solutions like Pymetrics or PredictiveHire not only match people to a job but also eliminate gender, racial and other biases.

Yet AI still can’t fully replace a human in the hiring process. Many studies show that technical skills often matter less than soft skills. Sociability, curiosity, and critical thinking are as important as tech expertise for 92% of executives. Another survey reports that when they are hiring, organizations with the highest financial return pay just as much attention to a candidate’s psychological traits, such as the ability to learn quickly, a sense of purpose and ambition. All of these features are much harder or even impossible to evaluate using existing math algorithms and pure logic.

To sufficiently uncover soft skills and emotional characteristics, AI requires self-awareness, which implies reflecting on one’s own thoughts. Most AI researchers believe that machines will learn to ‘know themselves’ in the way people do somewhere between 2036 and 2060. Until that time, we won’t be able to do without live recruiters when performing final candidate assessment.

Just-in-time learning and professional development

Millennials, or people born between 1981 and 1996, are already dominating the job market, and by 2030, they will comprise over two thirds of the US workforce. 87% of them underscore the importance of career growth and professional development. And what about the present day? The 2018 Workplace Learning Report shows that people strive for more flexible and faster upskilling, with 68% of workers preferring to learn at work and 58% opting for learning at their own pace.

To attract and retain talent, the most farseeing organizations already offer AI-powered learning opportunities to their staff. Most modern platforms for staff development exploit algorithms similar to those used by Netflix, Spotify, and other content-streaming services. They aggregate data about employees’ preferences, background, and skills to provide articles, courses, videos, books, and other resources which fit their needs best.

AI-fueled solutions like EdCast, Degreed, BetterUp or Axonify help businesses match workers with relevant learning content, design personalized development strategies, and build skills that really matter both for the company and employees. Moreover, with smart technologies, the learning process becomes as engrossing as watching ‘Game of Thrones’ (or whichever TV series you like best).

Identifying areas of stress and preventing toxic behaviors

Another promising usage of AI is to control fraud and non-compliance. Intelligent software can filter email traffic, and perform sentiment analysis of comments made by employees. In such a way, it can recognize patterns indicating ethical lapses, workplace bullying, loss of motivation, or fraudulent behavior. This information will allow HR managers to address risks before problems really occur.

One of the predictive tools already existing in the market is KeenCorp. To detect risk areas, it analyses emails, chats in messaging systems, and other digitally written texts created across the organization over the past two years. Then the system starts aggregating and evaluating real-time conversational patterns. As a result, managers gain insights into what is actually happening in the organization at the personal level, whether there are any areas of tension and conflict, and more.

Finding hidden stars and rising leaders, and managing talents

If AI is able to identify possible fraudsters, can’t it find high-potential leaders, hidden stars, and influencers inside the organization as well? The fact is it can, or at least is already taking steps in this direction. There are hundreds of psychological studies that list the personality traits of effective leaders. Here, the AI mission is to help HR managers spot desired characteristics and behavioral patterns in existing employees.

Again, an organization’s real-time communication flows become a source of valuable information. By measuring interactions, AI solutions like TrustSphere can answer the following questions:

  • How do employees collaborate with each other as well as with clients and suppliers?
  • Who influences whom in your company?
  • Who acts as a leader?
  • Which collaboration network is the most productive?
  • What small changes in the behavior would lead to better performance?

HR managers can use the results of internal communication audits to improve employee engagement, align team members with organizational goals, give ambitious people more opportunities to shine, retain talent and predict employee leave. The opportunities are endless, though a human perspective is needed to explore all of them.

The story of success to be continued

We must admit that: AI is not omnipotent. It lacks the intuition and imagination of a person. Its decisions are based on pure logic while people are often irrational, and HR managers shouldn’t ignore this fact. Even the smartest chatbot is not ready to take the place of a human being in circumstances such as conducting a final job interview or preventing a vital employee from leaving their job.

What AI-fueled tools can really do best is to aggregate data, recognize and match patterns, and make data-driven predictions. Algorithms are able to analyze millions of pieces of information in seconds. Yet it’s still up to a human expert to evaluate the results and make the final decision.

In order to stay competitive under current market conditions, AI is inevitable. In the future, AI systems will become even more clever, reliable and specialized. The story of cooperation between artificial intelligence and human resources is only beginning, and we’re sure that new breakthroughs are just around the corner.

<a href=""></a>