Computers will never fully replace humans in hiring

blog 03.jpg

The latest wave of data aggregate services in the recruitment space is an indication that the world of hiring is changing and changing in a way that will help companies find and choose a higher percentage of employees that fit into their corporate environment. Where these services are certainly a tool that Hiring Managers and other professional Recruiting Services can utilize, companies can ill afford to rely exclusively on these tools.

Like a hammer, these tools perform a function, but it’s the human element that determines their efficacy, much in the same way it takes a human to determine the strength with which to pound in that nail.

An article written in the New York Times, by Claire Ann Miller, called Can an Algorithm Hire Better Than a Human?, covers this subject well and concludes that a mix of both a data aggregate service and a human with judgement is where the future lies.  She states, “Data is just one tool for recruiters to use…”.

The article makes a very good point that algorithms can ignore the biases that commonly befall the recruitment process.  How well the recruiter gets long with the prospect, should have no bearing on their qualifications for a job in a different department.  What college they went to, and where they grew up or were born are rarely serious considerations for most jobs. And so it is true as well, with the more “dangerous” biases such a race, religion and sex.

The missing point of the article is that algorithms succeed at ignoring these biases not because they are algorithms and can’t “think and feel” these which is the argument, but in fact it is because the algorithms are separate from the corporate structure they are able to make recommendations based on performance and past history.  Again, it’s because these algorithms are separate from the business that they succeed in offering applicants and therefore can’t fall into the trap of calling biases “a difference in business culture” when it’s plain racism or perhaps not hiring a qualified woman because she is a woman, and due to an influence of “gender inequality” within the rank and file of the company.

Lest one envision a world where you type in your position requirements and two minutes later you receive three perfect applicants to your email inbox, consider the following aspects:

  1. Data aggregate services need to be programmed by humans, and humans that have experience in recruitment.
  2. Data aggregate services can only rely on data that is available to their algorithm both in terms of the applicant and your company.
  3. Data aggregate services work with data from the past and cannot know the direction and future of your company.

Let’s take each point, and expand upon them.

Data aggregate services need to be programmed by humans, humans who have experience in recruitment.

Humans posses the ability of judgement. They can determine relative importance between requirements and weigh an overall value of the person versus perhaps the time it may take to train them in a specific skill they may be missing. The right person hired, may not get the job when a month of training, would complete their skill set.  This takes judgement and the complexity of being able to program that sort of judgement into aggregate data services is still out of reach.

Data aggregate services rely on data that is available to their algorithm both in terms of the applicant and your company.

The elephant in the room regarding all these aggregate services, is that they are as reliable and accurate as the data fed. Aggregating data from Social Media and other places in the public domain can and does provide a methodology on where to start any prospective search, but the limitations to such a system become evident when considering that corporations and individuals rarely put forth in public domain information that would entail the full reality of the soft skills needed to work there. Websites such as Glassdoor, provide a forum for employees to list their experiences of working in any specific corporate environment and are a promise of opening the doors to the “real” inside scoop.  However, without validated and without sufficiently large datasets, these types of sites may (and probably do) skew the data.

Data aggregate services work with data from the past and cannot know the direction and future of your company.

The other facet of data aggregate services that must be taken into consideration, is that they work with datasets of information of the past of the corporation and of the individual.  Though the past is often an index and a part of analyzing the prospective applicant, a person tasked with hiring new employees would be advised to consider how new changes in management invariably change the corporate environment or how changes in market conditions can also dictate the pace of the company therefore changing and rendering less important various factors of corporate culture of the past.

As to an individual past, it suffices to point out, that almost anyone knows stories and examples of individuals getting a position and finally succeeding where before he was never a good fit, and reversely that candidate who has the perfect qualifications and history, failing miserably in his new environment.

In summary, new hiring data aggregate services offer a promise of being able to offer qualified potential recruits.  If one takes into consideration that the candidate or position would encompass a person generally disposed to being on the internet, using Social Media and otherwise not a private person, these can help in the search.  But algorithmic based computer logic will never replace the human element that encompasses such qualities as intuition, judgement and the ability to factor in the intangibles.

Professional recruitment firms, like my own, in Hiring End Hiring, have years of accumulated knowledge that helps choose applicants with the proper soft skills, experience and other qualifications.  We have an 85% successful placement and long-term hiring rate based on this experience and to which no robot can hope to replace.

Charles Harris