Automated hiring systems prevent millions of good candidates from getting through the...

nanoguy

Posts: 917   +12
Staff member
The big picture: For years, companies have been trying to optimize the hiring process by automating as much of it as possible, but in doing so they've come to a point where a lot of viable job candidates get rejected because of extensive use of automated hiring software, which is considered convenient and less biased than humans. But while most hiring managers acknowledge the issue, the reliance on screening software is so high that many aren't willing to look for alternative ways to solve the problem.

Automation is changing our lives in more ways than we can imagine, but sometimes it fails to deliver on its many promises. At a time when companies are struggling to find skilled workers, many potential candidates never seem to make it past the first stage of the hiring process -- getting their application of CV selected from an immense pile.

According to a study published by the Harvard Business School, employers may have pushed automation too far in the quest to process job applications faster and more efficiently.

The software systems that do things like sourcing and filtering candidates, managing the application process, performing background checks, and scheduling interviews can be of great help, but it also makes millions of potential candidates invisible to recruiters.

There are, of course, other factors that make it difficult for people to find employment, but the study explains that automated hiring software powered by machine learning and natural language processing has the biggest impact by far. No less than 75 percent of US employers use these systems, compared to 54 percent and 58 percent, respectively, in countries like Germany and the UK. In the case of Fortune 500 companies, researchers found that practically all of them use automation extensively in their hiring process.

Lead author Joseph Fuller told the Wall Street Journal that some of these automated systems often filter out candidates based on poorly-chosen, broad, or irrelevant criteria, and typically exclude from consideration those who would only need some additional training to fit the desired role. Examples include hospitals that look for references to computer science in the CVs of candidates for a nursing or data entry job, and retail stores that reject people who apply for a store clerk position but don't list "floor-buffing" as one of their skills.

A wide majority of employers -- 88 percent -- acknowledge the issue, but most wouldn't go back to a more traditional recruiting process. Instead, they believe the automated systems can be improved over time. A notable exception is Amazon, who has hired over 450,000 people since the start of the pandemic, and initially ran into issues with automated hiring software. In the meantime, the global recruitment technology market is set to reach $3.1 billion by 2025, almost double the size it had just a few years ago, so this is the new reality -- one where you can't get through the front door at a company because an algorithm and not a human recruiter decided you were not a good fit.

Permalink to story.

 

Squid Surprise

Posts: 4,151   +3,330
So basically... companies are simply configuring the AI poorly...

This isn't the fault of the automated software, but the individual companies that are using it.

It would be foolish to quit using automation just because of this - and the article even states that most companies don't feel they have to - just improve the software (and filtering system) and it will still be vastly superior to human recruiting.
 

Sausagemeat

Posts: 689   +487
I’ve been affected by this. I’m a dual national British American but if I put down British qualifications on the application I never hear anything back. All the work I have had in the USA is through word of mouth rather than from an application. I did get to speak to a recruiter who told me the system would reject my resume (CV) as it doesn’t recognise the British qualifications I hold.
 

kiwigraeme

Posts: 481   +368
Should just be a numbers game - ie throw out obvious red flags .
It's hard to do good post analysis - if 100 people apply for the job - you should be able to employ someone in the top 50 - then congratulate yourself how good you are . Plus people are bias and prone to failure - Why was electric shock treatment used broadly for a long time - the proponents BS how great is was for over 30 years - in reality it has a very narrow case use - These Egos in charge had PhDs and all that stuff .
In all the years of employing for my small business - the CV was the last thing I looked at ( TBH I did not need that much experience or skills - I could train people )

Number 1 - was the cover letter - You know those sad stories of people saying I applied for 500 jobs and heard nothing back - yeah mass sending generic stuff Does Not Work - I did reply , even replied most of times to people from sub -continent with no visa - never got a thanks from people without visas- Note a rejection letter can be an opportunity - if you write back with a thanks and ask what , why how in a nice way - it impresses people and shows character - can not be whiny or negative .

Number 2 - The Character of the person

That's it - right attitude, good character , good presentation - a bad employee is a liability .

Americans are very good at selling themselves - so need to be vigilant :)
They will tell you what you think you want to hear and present friendly and well.

I never had problem getting a job -2 examples.
When I was 15 or 16 rang up a cherry orchard to see if needed pickers - Guy said I take your number and let you know . I said I turn up tomorrow and if you don't need me send me home.
Got up 6am - cycled 7 miles on old one speed to orchard ( you pick early before heat ) . Turned up - handed me picking stuff gave me a row - youngest and one of the best pickers ( my family had 15 plum trees to earn some extra money- so already was fast - was contract anyway - but I also picked without damage nor missing harder to pick ones ).
2 First real job out of Uni - interview with Deloittes
About to go out door in new suit - just moved house - flatmates dog runs off . I phone as say I will be late and gave reason . Looking back as an employer now - I'd be impressed a young person could make the correct decision - as some stress involved . Means they won't hide serious mistakes etc .
Want to get a job - change your interview time to be last to be interviewed - Silly people who want to see 50 people give up remembering and quite often take the last good person they remember.
 

brucek

Posts: 863   +1,246
In an environment where applicants were plentiful and jobs were getting well-filled, I could see how fixing this would not rise to a top priority.

I'm less able to believe that in the current environment, with some employers desperate for candidates, that they are going to let this sit unaddressed for much longer (if they haven't already been working on it.)

Either way, people have been making hiring choices on all sorts of low-quality data and low-quality reasoning for a long time, and I don't see that part of the big picture ever changing.
 

Mr Majestyk

Posts: 824   +730
So basically... companies are simply configuring the AI poorly...

This isn't the fault of the automated software, but the individual companies that are using it.

It would be foolish to quit using automation just because of this - and the article even states that most companies don't feel they have to - just improve the software (and filtering system) and it will still be vastly superior to human recruiting.

AI is supposed to learn, it doesn't look like it's got any smarts at all, is just dumb old code using a check list of filters. Software rejects 75% of candidates up front and bosses complain no one is coming in for interviews, admit it's a problem, refuse to do anything about it, but continue to b!tch and moan.

You couldn't make this level of incompetence up, but hey if HR is involved there's your first problem.
 

Norsiiii

Posts: 80   +105
Makes no difference to the business - if they get 30 applications for a position, and 10 of them would do a great job, but the system 'wrongly filters out 5 of them' the firm will simply end up with one of the other 5 great candidates. Who cares, they still got a good person for the job?

To suggest that firms should be adjusting their automated systems to filter out fewer candidates automatically is contrary to the very purpose for which these systems are intended - to filter out as many candidates as possible with the least amount of input from humans required.

What's the difference between manually whittling down 10 great candidates to 1 person you offer the job to vs automatically throwing out 5 great candidates and manually selecting 1 of the remaining 5? There is no difference, except for the human time cost.

Changing the automated selection criteria is not going to increase the number of available jobs, it IS a zero sum game
 

brucek

Posts: 863   +1,246
Makes no difference to the business - if they get 30 applications for a position, and 10 of them would do a great job, but the system 'wrongly filters out 5 of them' the firm will simply end up with one of the other 5 great candidates. Who cares, they still got a good person for the job?
It starts making a pretty big difference when the software can't identify any matches in the pack on its own, and the company no longer has any other good means to find the best-available-even-if-not-perfect matches. It's an interesting issue but as you point out, after however long it takes to adapt, we'll ultimately be back to hiring one person for one job.
 

nodfor

Posts: 92   +159
t.

Changing the automated selection criteria is not going to increase the number of available jobs, it IS a zero sum game
No it isn't a zero sum game.
It is hard to determine what sum game it is, but it is unlikely to be zero sum.
You get an excellent candidate for a position - great. But what if you get a mediocre one? Maybe he won't be able to carry all the weight himself, so u have to hire another one or maybe even 2. The firm loses money due to extra wages but unemployment goes lower.

And what if you lose candidates that could turn your business around, it remains stagnant and eventually goes belly up? Then tens of jobs could be lost.

In a competitive world though, my guess is scenario number two is more likely to happen. If an economy systematically fails to hire the most qualified candidates, in the long run said economy will suffer and jobs will be lost.

 

Norsiiii

Posts: 80   +105
It starts making a pretty big difference when the software can't identify any matches in the pack on its own, and the company no longer has any other good means to find the best-available-even-if-not-perfect matches. It's an interesting issue but as you point out, after however long it takes to adapt, we'll ultimately be back to hiring one person for one job.
If a firm is getting dozens and dozens of responses to an open position and they're noticing that their automated system is filtering out every single one of them, it would quite quickly become apparent that there's an issue and they're not exactly going to just press on with using it as is

No it isn't a zero sum game.
It is hard to determine what sum game it is, but it is unlikely to be zero sum.
You get an excellent candidate for a position - great. But what if you get a mediocre one? Maybe he won't be able to carry all the weight himself, so u have to hire another one or maybe even 2. The firm loses money due to extra wages but unemployment goes lower.

And what if you lose candidates that could turn your business around, it remains stagnant and eventually goes belly up? Then tens of jobs could be lost.

In a competitive world though, my guess is scenario number two is more likely to happen. If an economy systematically fails to hire the most qualified candidates, in the long run said economy will suffer and jobs will be lost.
If you're giving jobs to mediocre candidates just because the automated system let them through it's initial filtering then you didn't do your final manual interviews/evaluations property, and it's entirely your own human fault for giving them the job offer.

Further, if you're talking about any sort of higher-up positions that would be likely to have high-level whole-company-trajectory sort of influence then you would 100% be selecting the candidate manually. Nobody is uploading job ads for CEOs or upper management onto Indeed. Similarly, no graduate employee is going to make or break your company (and if they are, then you've got much bigger problems to worry about).
 

wujj123456

Posts: 58   +36
TechSpot Elite
The other questions is how the automation is doing relative to human otherwise would. It's not that human don't make mistakes. If recruiters would otherwise get drown in the sea of applications, have to apply more ad hoc or inconsistent criteria, missing out even more good candidates, then it's not hard to understand why the companies still pick automation.

The value of the study to me is calling out the clear need for companies (both the vendor and employer using it) to validate its results regularly, analyze the patterns and correct any issues to reduce missed opportunity. If you are having problem filling a position, perhaps check if anything gets dropped out along the pipeline unnecessarily and perhaps give a shot to some "unqualified" candidates to validate the results. This might sound obvious, but in a data-driven world, a study confirming its potential is a good reminder.
 

Dimitriid

Posts: 953   +1,766
I'd go one step further: the AI problem is very likely just exacerbating problems that are already present in your average HR department and their hiring process: most of the high level (Managerial or executive level people, not just guys actually looking for candidates) HR people are well, self-important jerks: The other departments would give them very specific, simple requirements like say "We need someone who can code in X and we prefer if he has experience in Y type projects"

If you would go just by what the actual people that need the employee in their department actually need, each position would net them thousands and thousands of candidates from day one most of the time.

But it's HR that's starts first of imposing their own limitations like "This is an intermediate position, so our policy is that we require at least a Bachelor's degree and 5 years of previous work experience" and suddenly, your list went from pretty much most of the IT people that now how to code X applied for Y applications down to a couple hundred at best.

Then HR takes it a step further and start vetting with their very own biases based on utter nonsense that has no place in reality like "This candidate wasn't enough of a 'self starter' in our screening interview" or "This candidate didn't immediately responded to our email which shows a lack of drive" and so on.

It's really not that hard for mos people to find "Guy who can code X language and knows it's way around Y applications" and hiring process go from checking out *all* guys and going with the guy who probably would expect a salary at around what you're willing to pay according to their experience and make it hard requirements and add their own, pointless requirements and just vet themselves out of candidates or teach candidates how to cheat the system and target specific jobs effectively perpetuating this cycle.
 

Xex360

Posts: 141   +181
I'd go one step further: the AI problem is very likely just exacerbating problems that are already present in your average HR department and their hiring process: most of the high level (Managerial or executive level people, not just guys actually looking for candidates) HR people are well, self-important jerks: The other departments would give them very specific, simple requirements like say "We need someone who can code in X and we prefer if he has experience in Y type projects"

If you would go just by what the actual people that need the employee in their department actually need, each position would net them thousands and thousands of candidates from day one most of the time.

But it's HR that's starts first of imposing their own limitations like "This is an intermediate position, so our policy is that we require at least a Bachelor's degree and 5 years of previous work experience" and suddenly, your list went from pretty much most of the IT people that now how to code X applied for Y applications down to a couple hundred at best.

Then HR takes it a step further and start vetting with their very own biases based on utter nonsense that has no place in reality like "This candidate wasn't enough of a 'self starter' in our screening interview" or "This candidate didn't immediately responded to our email which shows a lack of drive" and so on.

It's really not that hard for mos people to find "Guy who can code X language and knows it's way around Y applications" and hiring process go from checking out *all* guys and going with the guy who probably would expect a salary at around what you're willing to pay according to their experience and make it hard requirements and add their own, pointless requirements and just vet themselves out of candidates or teach candidates how to cheat the system and target specific jobs effectively perpetuating this cycle.
Indeed, but I wouldn't blame HR only, when you read job descriptions it's pathetic, they want a someone holding a PhD in a variety of fields with multiple years of experience, while the job itself requires nothing special.
HR add they layer of stupidity like not answering fast enough, or not answering as they expect them to.
 

Dimitriid

Posts: 953   +1,766
Indeed, but I wouldn't blame HR only, when you read job descriptions it's pathetic, they want a someone holding a PhD in a variety of fields with multiple years of experience, while the job itself requires nothing special.
HR add they layer of stupidity like not answering fast enough, or not answering as they expect them to.
Fair enough that's why I tried to made the distinction between regular HR which are mostly just low level employees doing the calling, collecting the paperwork, checking the sites and emails, etc. And the HR executives and other executives that are responsible for actually dictating the company policies which create all these layers of nonsense requirements and complications.

In fact I'm sure most HR personnel in charge of actually trying to find candidates are just as frustrated, if not more so, with the policies they're told to enforce and the people they're told they have to turn away because of dumb stuff.
 

James Ryan

Posts: 9   +10
I've jumped through a few of these systems in Australia. You'd be surprised how many companies use them now. After the first 2, I decided to skip all job applications that had high level automation. At least automation that was visible from outside.
 

Puiu

Posts: 4,864   +3,749
TechSpot Elite
Automated systems are not good for hiring people. Only after talking to someone can you actually understand if he's fit or not for the job. The best CV doesn't make for the best candidate.

You don't have to automate absolutely everything...
 

Irata

Posts: 1,675   +2,808
I believe the point of automated selection, standardized templates, assessments etc is to shirk personal responsibility.

If HR do selections based on their judgement, they can be held accountable if it does not turn out well. An automated system did it ? Not their fault.

This may not be bad in a situation where you have 1000 applicants for the same job and requirements are not specific but this does backfire in situations where there are very few applicants and they can take their pick between several positions.
 

nodfor

Posts: 92   +159
Similarly, no graduate employee is going to make or break your company (and if they are, then you've got much bigger problems to worry about).
If your hiring process is flawed, it won't affect a single hiring but most of them, so it will most likely have an impact
 
Last edited:

James Ryan

Posts: 9   +10
Well when crap workers are what they really want. I see no need in wasting our time.
It all comes down to cost and knowing that they have an endless pile of people to consume. They just wear them out and move on to the next. Regardless of the skill level. There's always someone ready to fill the slot.
 

Xex360

Posts: 141   +181
Fair enough that's why I tried to made the distinction between regular HR which are mostly just low level employees doing the calling, collecting the paperwork, checking the sites and emails, etc. And the HR executives and other executives that are responsible for actually dictating the company policies which create all these layers of nonsense requirements and complications.

In fact I'm sure most HR personnel in charge of actually trying to find candidates are just as frustrated, if not more so, with the policies they're told to enforce and the people they're told they have to turn away because of dumb stuff.
Executives tend to be the dumbest people in the company. Fortunately for them their salaries are negatively correlated to their IQ.
 

sac39507

Posts: 394   +211
But for some reason, the worst candidates seem to keep filtering through and getting hired. I guess it pays to be dumb. The "Algorithm" picks this up. Must be the same coders who worked on election systems.
 

Bp968

Posts: 234   +165
Next level is automated cv generators so you better fit the automated hiring systems.

This. This was my first thought. If you automate something as vital as a human beings access to a livelihood they *will* fight back and being a capitalist society someone will make a job and business out of gaming those automated systems.

The AI might learn but the best it can do is follow a set of instructions (we want candidates with x qualifications, etc). It will become an arms race and everyone will lose (except the folks making and selling the software).