Want to make your LinkedIn profile stand out in 2023?
You may know that LinkedIn is a compelling way for you to get noticed when looking for a new job. LinkedIn is a platform that
Connor Gwilliam is a recruiter within the Engineering department who specialises in placing data centre engineers with leading FM companies.
Artificial intelligence (AI) is quickly becoming a part of our everyday lives.
It suggests what we should be watching or listening to next on our streaming services, it shapes our social media feed and offers us personalised adverts tailored to our search history. Even the facial recognition function used to unlock your smart phone is AI in action.
The endless number of applications for artificial intelligence hasn’t stopped short of the recruitment industry either.
Some companies out there are already using it to refine the hiring process for their clients. And others are leveraging the functionality of ChatGPT for increased productivity and quality of writing. It can even be used to find passive recruitment targets or to suggest where to post job adverts to attract more suitable candidates.
There are certainly plus points for AI and machine learning in recruitment. AI can:
AI’s potential to streamline the recruitment process isn’t its only selling point. One area where artificial intelligence supposedly excels is that as it’s computer-driven, it can’t introduce bias into the recruitment process. Candidates can be shortlisted using untimed screening programs without the need for CVs, checking social media profiles or using any kind of demographic data.
AI can filter CVs quickly using keyword analysis, automatic reference checks, and social media analysis. For positions that attract a lot of interest, this AI pre-screening can save a considerable amount of time for recruiters which will ultimately reduce costs. In fact, AI could be so effective that it could replace 16% of recruitment sector jobs by 2029, according to a 2019 report.
And it doesn’t just filter candidates by job history and qualifications. Some specialist AI recruitment companies offer online candidate testing which also scopes out their soft skills. Another company has perfected an AI-based system that uses one-way video interviews to collect additional information about candidates. The audio is converted into text and sent to the hiring company, which then decides who progresses onto an interview with a real person.
One area where artificial intelligence supposedly excels is that as it’s computer-driven, it can’t introduce bias into the recruitment process. Candidates can be shortlisted using untimed screening programmes without the need for CVs, checking social media profiles or using any kind of demographic data.
The truth is that we still have a lot to learn about AI and machine learning, and how to effectively use it in the hiring process.
In 2018, Amazon.com scrapped its AI recruitment tool as it showed a bias against women. The AI system had been introduced in 2014 to review applicant CVs, giving each candidate a score from 1-5 stars – similar to how products are ranked on its website. However, by 2015, it had become clear that the ratings for software developer and technical jobs were not being viewed by the AI in a gender-neutral way.
This was because the AI system had learnt to vet candidates based on a decade of CVs submitted to Amazon. And with big tech being a male-dominated industry, most of the previous applicants had been male. So the AI taught itself that male candidates were a better option than female ones.
The potential for statistical bias is a difficult nut to crack for machine learning. Essentially, the models created by the computer are only as unbiased as the data provided by its human operators. If a position is statistically held by people who are mostly male or have been to particular education institutions, the computer will seek them out at the expense of potentially better qualified or more suitable candidates.
AI programmes can also be unfairly biased against candidates who don’t speak English as a first language. Most AI’s recognition of spoken words is based on standard English as spoken by native speakers, so AI programmes can struggle to understand the subtleties of language and tone of voice of non-English speakers thereby introducing the risk of racial bias.
For companies proactively looking to hire a diverse range of employees, any inherent bias embedded in machine learning could make it difficult for AI to deliver the rich, diverse pool of talent they want to be able to choose from.
In years to come there may be legislation which better regulates AI-driven recruitment practices which will help alleviate some or all of these issues. In the US, New York State recently introduced legislation to prohibit the use of AI tools in recruitment unless it passes a bias audit. The results of the bias audit must also be published.
Closer to home, the European Commission is also looking at regulating the AI industry, while in the UK a white paper is expected to be published this year as part of the Government’s National AI Strategy.
Chatbots and screening programmes which filter CVs make their judgements based on past hiring decisions. This black and white approach quickly filters out people with lower educational achievements, different skills, and experience, and those with gaps in their work history.
In the process though, it filters out good candidates who may have had time off to raise children, or who didn’t finish university because they couldn’t afford the fees. It also weeds out people who are looking to make a step into a new career with transferrable skills from another industry. People that would have been otherwise selected to progress had they had a person-to-person interaction, rather than CV vs AI.
AI software can be used to analyse soft skills, such as teamwork (using we more than I in responses for example) but what it can’t yet comprehend is cultural fit. Is this person going to fit in with the personalities in the team, do their values reflect the companies? These are the nuances in the recruitment process that only people can understand and quantify.
AI definitely has a role to play in the recruitment industry. Particularly to automate laborious tasks saving on the time and expense of filling a vacancy. But it’s no substitute for the specialist industry knowledge, and the client and candidate relationship building that real, human recruiters offer. You can’t teach a machine to understand people. Everyone’s different. We’re so much more than a statistic or a piece of data to be processed. Some things will always require a human touch.
You may know that LinkedIn is a compelling way for you to get noticed when looking for a new job. LinkedIn is a platform that
As workplace management has evolved rapidly over the last few years, the facilities management industry and those working within it are challenged to keep up
We’ve seen an explosion in the number of clients looking for data centre engineers over the past 12-18 months, unsurprising as data centres are one