Artificial Intelligence (AI) is changing virtually every industry, corporate recruiting being no exemption. Primarily, AI will change the recruiter’s role through augmented intelligence, allowing recruiters to outsource some of the tedious or menial tasks of their job and focus on providing high-value results. The recruitment process’s intangible and abstract aspects, such as determining culture fit and developing relationships with hiring managers, will always be in the hands of the human recruiter.
Recruiting strategies that employ AI see:
- 35% decrease in turnover*
- 20% increase in performance*
- 4% increase in revenue per employee*
It’s not humans or AI, it’s humans and AI
Like many of the jobs that were thought to be entirely replaced by AI, automation, or robots (remember the Luddite riots in the early 1800s that destroyed textile machines as a protest against automation?), the future of recruiting will look much more like AI enhancing what humans do best rather than replacing them altogether. With AI, many recruiters will find that their role’s trivial tasks are no longer necessary and can deliver even better outcomes in the hiring and recruiting experience.
Many companies are already using some recruiting AI software for:
- Resume screening
- Digitized interviews
Recruiting for Tomorrow
The best mix is humans and AI working together. AI can, and should, handle repetitive tasks that ultimately don’t bring the hiring companies more value, while humans will continue building relationships and making decisions on unquantifiable inputs. At GRN Mid-Cities, we know that hiring companies aren’t as much concerned with the how as with the final result: quality candidates that fill a needed role. That’s why our focus is consistently improving where we can and employing technology so long as it makes sense for our promised deliverables.
There are positives and negatives with the current automation technology available for recruiting. This article will discuss the advantages, disadvantages, and how some companies are using progressive tech.
Advantages of Recruiting Automation
The recruiting process can include many repetitive, high-volume tasks such as resume screening. There could be thousands of applications for any single hire, and, historically, 75-88% of the resumes received are from unqualified candidates**. Even at 15 seconds per resume, that’s a lot of time spent looking at someone’s well-designed recycling bin filling. AI stands to improve recruitment by reducing or eliminating many of those tasks and saving time for the recruiter, hiree, and hiring company.
According to one source, AI saves their recruiters approximately 23 hours per hire*.
Reduce or Eliminate Bias
Applications filtered through AI could have some of their bias-identifying information hidden when presented to the human audience. Of course, none of us want to believe that we have biases, but the statistics have proven otherwise. Whether conscious or subconscious, these biases exist in humans and prevent workplaces from increasing diversity (which has been shown to be positively correlated to improved productivity and creativity).
For example, AI can hide the name, age, or race of applicants, so those pieces of information don’t impact the hiring decision.
Writing and Technical Aspects of Recruitment
Historically, hiring managers or recruiters write the job description, and while they may have a deep understanding of the job requirements, they are often not skilled writers. In the age of keywords, AI can write job posts that appeal to the specific types of candidates the company intends to hire. This modification saves time and makes the job description more accurate; plus, not writing any more job descriptions is definitely a win.
Additionally, some jobs include technical assessments during the interview process that could be reviewed and checked by AI. These are incredibly helpful for employment in accounting, sales, or engineering because they prove that the applicant has the knowledge basis to perform the job duties. Having a technical expert review the assessments when the results are often easily quantifiable creates a time-consuming task that is not very valuable. Having AI administer and check these sorts of assessments to qualify applicants is a significant advantage.
Matching a candidate to an open position and vice versa can be highly complex. There are many factors to consider on both the employer side and the prospective employee side. Cross-referencing data on the employer, position, and prospective employee, AI can identify the best-suited match based on tons of data points.
Disadvantages of Recruiting Automation
AI Requires a Ton of Data
A critical mass of data must exist for machine learning algorithms to understand the parameters and desired outcomes. Having a small data set can result in availability bias (the idea that features from a small sample that is readily available can be extrapolated onto the entire lot) that distorts the AI’s algorithm. This data necessity is required on both the employer and employee side.
AI Can Learn Human Biases
For all its praise about being an impartial judge, AI can quickly learn from the existing human biases and reinforce them. Historical data and results inform machine learning algorithms. If a large company has a history of prejudice in its hiring decisions, the AI will discover, learn, and employ it as well. This disadvantage of AI is particularly pernicious since we tend to put a little more faith in AI as being unbiased than we probably should. As they say in the world of data: garbage in, garbage out.
Skepticism from Decision-Makers
With the advantages and disadvantages of AI that we’ve already discussed in this article, it’s understandable that decision-makers would be skeptical. However, the truth is: they just don’t think AI is as good as humans at recruiting. While some of that sentiment is just the old guard thinking their way is the best way forever, AI still has a long way to go before Alexa is finding us new jobs.
Is the Science there for some of these AI algorithms?
AI is built with technology that exists today on the available science. One key idea that informs some AI algorithms is that of “microexpression,” the concept that tiny facial movements indicate unspoken emotions like fear, anxiety, or joy. While this sounds intuitive, the microexpression idea is not founded on exact or wholly supported science.
Additionally, some aspects of an AI interaction may not reveal a candidate’s best or most accurate self. A digitized interview certainly saves time and resources, but how a person interacts with a virtual interviewer could be utterly different than in-person. The data behind some of these features remains to be seen.