If you haven't already heard last night a computer program called "Watson" started a three night run as a contestant on Jeopardy. A "thinking" computer that takes just seconds to understand questions and come up with correct responses on a wide variety of topics, does this mean your sourcing department could be replaced by computers sometime in the near future?
The results of the show so far have been mixed but as a software developer myself what is more interesting than how well or poorly the computer does is the demonstration to the general public of the gap between human thought and the "thinking" a computer does. For the computer it's not having access to information that is the problem, it is understanding the question that is the challenge.
It's a problem space that anyone associated with computers and programming is well aware of and it manifests itself in many ways including, the focus of this blog, searching and sourcing candidates for specific roles. It's why in spite of the many promises made of "magic bullet" software solutions that can take over your screening and sourcing these systems have not, and if the amount of computing power behind Watson is any indication, can not deliver.
We all know that while there are many benefits to a large database of candidates in an applicant tracking system it can also be a drawback when it comes to actually identifying the best candidates for specific roles. From a computer science perspective it's a challenge, be able to understand and quantify what a resume says, be able to understand and quantify what kind of candidate you are looking for and then matching the two together.
From the perspective of a user trying to search the system it's an excercise in frustration as we have to spend a lot of time being specific with what we are looking for while still getting mixed results. Mixed meaning that we either aren't finding all the candidates who might be suitable or we are finding too many candidates who aren't actually helpful to our needs. "Innovations" like "boolean searches" (which as an aside aren't really that innovative since they've always been a part of underlying database theory), have helped but there is still a lot of work required on your part to get actually usable search results.
So back to the headline does Watson mean that your ATS will soon be getting smarter then ever? Just drop in a job description and find the actually three best candidates in your system? No need for manual resume screening anymore?
Well. No.
If you have been following along you'll note that Watson took years of effort from multiple people, incredible amounts of hardware, is focused on one specific problem "domain" (aka playing jeopardy) and still is far from perfect. All of which would tend to show that the "holy grail" of fully automating the sourcing and screening process is still quite some time away.
It also seems worth mentioning at this point that all of even this relies solely on hard, quantifiable skills and experience. The realm of the job interview with it's appraisal of soft skills and analysis of how well (or not) a candidate will actually fit in to a role and organization isn't looking replacable for a very long time if ever.
At any rate it would seem that despite the advancements in AI that Watson represents the practical applications for it are still some time away and limited, at least to start, to certain types of problems. This technology or not it is still going to be very difficult for a computer to read and more importantly "understand" exactly what a job description or candidate CV says.
In many ways it's a reminder of how the valuations and assessments of people's skill sets, that we to a degree find easy, really are quite complex. Speaking from a diversity hiring angle it throws a light on the sort of hidden complexity we see all the time. We often talk about the need to make sure you are looking at everyone and it is exactly this sort of hidden complexity where subjective descisions, the kind that make you blind to some candidates, lie.
Some food for thought anyway.
No comments:
Post a Comment