Published on 4th December 2019
Automation, AI, Machine Learning - Anybody who has anything to do with technology, (which is the majority of us in this day and age) would have heard about the movement in AI, machine learning, and automation.
Is Automation a no brainer?
Although there are varying opinions on how innovative technological advancements will impact our day-to-day lives, it is evident that it will somewhat have an impact on all of us.
From what I have learned, the purpose of these new processes, technologies, and applications is to make things faster, more accurate, less time consuming, save resources, and ideally money. In a perfect world, all of these aspects would come to fruition within all implementations of automation, although this is unfortunately not realistic. Other than the technical downfalls of implementing automation, i.e. initial developer time, increase in tooling needs, complex analysis, debugging, deployment costs, there are also a number of issues that arise from minimising human input listed below that I'll explain in further detail.
“Biased? Who me!? No of course not!!” – that is how most of us would respond when approached with a topic of recruitment bias – although unconscious bias is a human trait. This is one of the key human impacts that automation aims to reduce in recruitment, although there is also the potential for technology to make these matters worse.
This was the case with Amazon, in 2014 they started using AI to review job applicants with the aim to identify top talent. Automation was a successful tool in a lot of Amazon’s internal processes, including e-commerce, warehousing, and pricing decisions, and it seemed obvious at the time that the recruitment process could harness these tools. The tool gave candidates a rating of 1 to 5 stars via AI, where the top 5 out of 100 resumes would be provided to a Hiring Manager prior to any human reviews. By 2015 it was clear that the system had a bias towards males for technical roles, this was a result of the computer model observing patterns in resumes submitted to the company and designing a process based on male-dominant applicants. Ultimately after no confirmation that this bias could be removed from the program, it was canned from consideration in the recruitment process.
Is automation a reliable tool for recruitment?
The automation of recruitment tasks can have a positive impact on day-to-day tasks and workflows by increasing productivity, accelerate time-to-fill, reduce cost-per-hire, and improve overall talent profiling. Although does this outweigh the benefits that a human assessment provides to the recruitment process?
Through harnessing human judgment, the recruiter can use their abilities in understanding a candidate on a personal level, opposed to information on a CV. This ensures that skills, experience, and character traits are aligned to a job, or if they aren’t perfectly aligned, then are there other aspects that make up for the gaps. Human judgment asses if a candidate has the culture fit and motivation to learn, which can be a lot more valuable than expertise in the desired skill or ideal experience.
So where to from here?
I think that it is inevitable that we will all learn to embrace automation, although we must make sure these technologies are tried and tested before we bring them into business-critical aspects such as recruitment. With the increasing need to eliminate bias in recruitment to better the experience for both clients and candidates, the potential for automation to have a negative impact is a risk that we can’t afford. Our candidates are more than just ticks and crosses and deserve to be seen not only for what they have achieved, but also what they are capable of achieving in the future.
If you would like to talk about your technology recruitment journey reach out for a chat here.