Where does AI fit in when it comes to your hiring process? It’s a good question and all the more relevant since AI has been radically changing how background checks traditionally work. Supposedly AI can improve a traditional background check by taking out the need for human analysts (AI-based judgment calls) and therefore increasing efficiency exponentially. But is there a cost of removing humans from the work of Human Resources?
What does AI do, again?
Terms like Artificial Intelligence and Machine Learning imply that a machine is created to make a comprehensive, human-like assessment of situations and adapt accordingly. The reality of coding automated systems is far more complex than this author can imagine, but in every instance where AI is applied to streamline a particular function, it is given specific goals to achieve. In some respects, it’s extremely helpful in reducing manual labor and increasing efficiency for mundane, repeatable tasks, i.e. scouring publicly available information about an identity. Without AI, our access to information and the navigation of it thereafter would be unwieldy, if not near impossible.
Why is that relevant?
AI made global business what it is today (or even possible, for that matter), but accompanying every innovation is a discussion on the ethical implications of such improvements. From an HR and recruiting perspective, the question becomes “What are the limits of AI when it comes to making judgment calls about candidates’ data and digital history?” or perhaps more specifically, "Is it ethical to base a hiring decision solely off of a “score” given to you by an algorithm?" The technology is certainly out there already, but that doesn’t necessarily make it right.
Automation without human analysis may be where the industry is headed, but that doesn’t necessarily benefit your hiring process. An algorithm may be able to identify the correct individual and spit out a bunch of potentially problematic content, but it cannot provide relevant context or understand the full weight of the creator's intent.
Human intelligence is still necessary.
To this end, it is still absolutely viable to use human analysts to comb through the relevant data an algorithm has sorted through. Would you ask Siri or Alexa who you should hire? Let's say an algorithm made to alert the FBI to imminent threats notes that you've been googling how to make an explosive. That sounds problematic, and artificial intelligence would be able to flag it as such. However, what AI cannot do is discern whether you are planning to actually set a bomb off or simply researching your next action-packed novel (You’re a regular Tom Clancy!). The same goes for social media background screening. An algorithm can find all manner of content based on pre-defined filters, but would it flag a post like this?
A piece of software that combs through social media accounts would not be able to decipher that this photo is inappropriate, demeaning to hard-working women in the workplace, or sexually explicit. A level of human curation is necessary to decipher that the intent of this photo and it’s accompanying comments are not innocent.
That’s why in the evolving and ever-growing world of data, we still need humans to discern what other individuals are doing, especially when their current and future employment prospects are in the balance. Human intelligence is still the most reliable way of aligning values and nothing gets lost in translation because an artificial being hasn’t made any out-of-balance judgment calls for you. It would take an AI army to objectively quantify all human intention, and there is not enough tech on the planet to completely emulate our instincts. It’s why we have phrases like “You are not the sum of your worst decisions,” or “Wait till you meet her, she’s not what she looks like on paper.”
Your decision to employ someone boils down to so much more than a score on a report. Yes, those reports are there to help influence your decision, but should never be taken so far as to make your hiring decisions for you. You receive and interpret more information about your colleagues and candidates in a single day than AI could in a year-- don't let the machines take over just yet.