When AI goes too farBy Ron Onur Aksoy
When it comes to new technology, I am more than just an enthusiast. After all, I spend my days leading a company that helps other companies define and build out IT infrastructure to meet their business goals. So, needless to say, I am all for advancements in technology.
However, a new study conducted at Harvard University used Artificial Intelligence (AI) and Machine Learning (ML) to create a completely different approach to human resources, which is calculating when a person is likely to leave the company. The new algorithms take into consideration everything from past job history, skillsets, and on-the-job patterns to determine the likelihood of a person quitting.
To give a bit of background, the team conducting the experiment stated that research determined two main reasons as to why people leave their jobs: turnover shocks and personal reasons. To clarify, turnover shocks are any event that may lead to a person contemplating whether or not they should stay with the organization. For instance, mergers or acquisitions where the corporate culture may change or represent a risk to job security, or leadership changes, and so on.
As for personal reasons, the researchers categorized this as the level of job embeddedness that pertains more to the social aspects of the work environment. For example, whether or not a person has friends and/or an integral social network at their place of work, or whether or not they feel as though the company represents their own personal values, skillsets, etc.
In all, the markers do make sense. It is what companies have done for years when conducting job interviews and exit interviews—all created to determine why people may or may not leave. And rightfully so given that employee turnover can be both costly and disruptive. Furthermore, attracting and keeping the best and brightest is good for business in a myriad of ways.
But the question becomes how far do we take AI before it impacts people and their livelihood? And though I am not arguing about the methodology of the study, I am arguing about the implications and how they can impact people and the companies that utilize such an approach.
As someone who employs many people, I am the first to say that our people come first. And in doing so, the ultimate outcome is understanding people on a highly personal level: enough to know that job history, skillsets, and on-the-job patterns are just the tip of the iceberg. Furthermore, if a company is managing its human resources appropriately, social aspects paired with an appreciation of skillssets should rarely come into play, mitigating risk of departures by continually addressing the human factor.
The greatest issue that arises when such technology is put into place is the loss of connection with the human factor. So much so that using such technology could very well be off-putting to potential and current employees. If the perception is that one’s own job stability is being continually measured and calculated by so-called machines, what human factors are then important to the employer?
Again, technology is wonderful, but implementing an almost Orwellian approach to human resources the “human” part of the equation disappears. In the end, people are inherently complex. And with that, managers and leaders of companies need to embrace that nature to ensure real connections are made.
Yes, there will always be people who wish to move on, it’s human nature. But subjecting people to an algorithm to qualify and quantify their perceived worth to the company and whether or not they should be marked as a so-called flight risk is the exact point when technology goes too far.