Why We Don’t Use AI To Search For CVs

When used properly, AI can help streamline the recruitment process, eliminate human error and bias and it never gets tired when checking its 200th CV of the morning. But it can also eliminate the human ability to read between the lines and identify potential.

How AI reads your CV

Some AI screening tools use keyword-based algorithms to screen for keywords, phrases, and patterns in the text to sort candidates, others use grammar-based machine learning algorithms that use a list of predefined grammatical rules, breaking down words and phrases on the CV to understand the meaning of each sentence against predefined job profile criteria.

So in the same way that spell checkers alert you to a typo, algorithms sift through your CV’s keywords and information about your work history. It is, at its most basic, merely a keyword algorithm.

These algorithms search for continuity of work history, job title progression and education and while different companies will use different metrics, the benefit of not having to search through a never-ending pile of CVs is obvious.

However, calling them basic keyword algorithms doesn’t do it justice. They are used by huge corporations around the world and unlike a simple algorithm, they can learn and evolve. Using thousands of previous versions of a job to correlate and scanning data on CVs to find the most appropriate candidate for the talent recruiter.

AI recruitment software

There are several off-the-shelf AI recruitment software available including:

Arya is a fully featured Artificial Intelligence Recruitment Software designed to serve Agencies, Enterprises. Arya provides end-to-end solutions such as Intelligence Driven Decisions, Resume Screening, Job Posting, Recruitment Analytics in one place.

Manatal is a cloud-based recruitment software designed for agencies and corporate HR departments. It is an AI recruitment software designed to streamline recruitment processes from sourcing to onboarding through digitalization and seamless integration to Linkedin and other job boards.

X0PA offers AI-enabled Software as a Service (SaaS) solutions that uses Machine Learning to score and rank talent identifying the best-fit candidates for each company and role. It enables scientific and objective choices in hiring and selections using its patented AI algorithms to make hiring efficient and bias-free. 

Paradox.ai uses conversational AI to help companies automate their recruitment process.

Phenom Intelligent Talent Experience is an AI recruiting platform that includes career site, chatbot, content management system, CRM, AI scheduling, video assessments, campaigns, university recruiting, talent marketplace, career pathing, gigs, mentoring and referrals.

SeekOut is a cloud-based recruiting solution designed to help businesses automate processes related to candidate hiring and combine talent acquisition and management into one platform. Built on human-driven, AI-assisted technology it  lets recruiting agencies receive candidate recommendations based on job descriptions.

HireVue uses AI to analyse candidate interviews and provide insights into their skills and personality traits. It boasts clients such as Unilever, Hilton, Vodafone and Oracle by engaging candidates all the way from first contact to hired.

Is AI recruitment software biased?

According to a 2019 study, “Without human intervention, it is easy for algorithms used in the recruiting process to reproduce bias from the real world”. 

AI algorithms on facebook advertising in the US was found to target specific demographics. Facebook ads for supermarket cashier positions were shown to an audience which comprised of 85% women. Meanwhile taxi jobs were targeted to an audience that was around 75% black.

AI resume scanning tools are not perfect and can have biases. For example, Amazon stopped using a hiring algorithm after finding it favored applicants based on words like “executed” or “captured” that were more commonly found on men’s resumes. Another source of bias is flawed data sampling, in which groups are over- or underrepresented in the training data. 

Are AI algorithms sexist or racist? No, but it does demonstrate how algorithms reproduce bias from the real world, without human intervention, just from being fed historical data sets. However, many of the flaws of AI are leading to corrections. IBM’s AI-based tool Watson screens resumes without considering personal details such as age or gender. It is clear that with time, AI-based technology will counter its own flaws and further improve recruitment processes.

Why we don’t use AI to search through CVs

Ascent People use human intuition for sourcing candidates because even the most sophisticated AI technology cannot replace a great recruiter. The best IT recruiters already know how to automate the tedious aspects of recruitment, but we also know how to find the top talent requires a bit of intuition, knowledge experienced and asking the right questions.

There are several reasons why IT recruitment agencies might not use AI to scan through CVs:

Unlikely to pick up on the major strengths of a candidate

Some AI algorithms are unlikely to pick up on the major strengths of a typical candidate who doesn’t meet the precise criteria given in the job description. These are candidates who have worked for competing companies, or used particular software or been active in comparative technologies. They may even have a job title that says one thing but are actually involved or responsible for something much more nuanced.

AI can reinforce inequality in job recruitment

Instead of reducing bias AI algorithms might actually enhance it. They are not biased because of a technical glitch or because of some sci-fi scenario in which the robot takes on a personality of its own, usually the responsibility lies with the people who input the selection rules. This means that if historical data sets include some hidden, random or incidental bias in the data set, the resulting current algorithm may demonstrate them too.

As outlined in the previous section, AI bias might not even have anything to do with the actual skills required for the role. Like school children with a dictionary, they doesn’t always learn the right words.

AI can overlook qualified candidates with underachieving resumes. 

It is surprising in this day and age, with the amount of information and guidance on the internet advising and showing candidates how to write a CV, that many of them still don’t do a very good job of it.

It’s all about reading between the lines, and this is where experience comes into play. A highly experienced IT recruiter will understand the particular skills being used and question those previously used in order to make sound judgements to their suitability. Or, quite simply, we know them and can ask the right questions that provide evidence of their skill-set and experience for the client requirement.

Sourcing top talent is about making connections 

Human recruiters are better at making those important connections with candidates and clients than any AI ever can. No AI recruitment software will call a candidate back to question inconsistencies in their CVs and root out their achievements hidden under grammatical errors.

Not all stakeholders are on the same page when hiring for a specific job role. AI isn’t going to have a conversation with the hiring manager, the IT director and the HR team to align job (or even project) descriptions. AI isn’t going to question the client to make sure they are actually advertising the key skills required. 

We ask those difficult questions and we have those awkward conversations.

Hard to fill roles

There is a reason why those roles are hard to fill; they require either the very best talent (talent that needs convincing your company is the best to work for) or because the skill set is so niche, a standard CV just won’t match up to it where no amount of AI can filter the ideal candidates.

There may be a shortage of candidates based on location issues, lack of flexibility, company culture or where there is negative reputation around a brand or company. Then there is the tricky question of ‘is it a competitive salary?’

If we used AI recruitment software for these roles, we would be wasting our time. We don’t, we use over 40 years combined experience to identify, negotiate and convince for the benefit of clients and candidates alike.

So, while AI can help automate some parts of recruitment, and help recruiters to improve process efficiency, it simply cannot replace human intuition and decision-making when it comes to selecting candidates for a job.