Employers and job candidates are dueling with AI in the hiring process

The process of getting a job is looking more and more like a battle of the bots. Candidates are increasingly using artificial intelligence software to trawl for openings, tailor cover letters to specific jobs, and even apply for jobs all on its own. On the other side of the battlefield, companies are upping their use of AI to find candidates, communicate with them, and rank their applications. For many candidates and employers AI has become table stakes, and the situation can feel a bit like an arms race to try and keep up. 

The question has become whether each side offloading parts of the hiring process to bots, so that essentially computers are talking to computers, will create more signal than noise, and undermine the entire process of recruitment altogether. 

As early as 2020, more than 90% of employers were using software (including, but not limited to, AI) to initially rank or filter mid-and high-skills candidates, according to a survey of more than 2,000 executives. Now it’s more like 100%, according to one of the report’s authors, Harvard Business School management professor Joseph Fuller. Currently about a quarter of HR departments are using AI, primarily in talent acquisition, according to a new report by the Society of Human Resource Management, which expects that to rise to 50% next year.

Top recruitment software and applicant tracking system iCIMS, which boasts 3 million users including a quarter of the Fortune 500, uses AI and generative AI to connect job seekers with pertinent roles and help recruiters write job descriptions, communicate with candidates, and rank applicants. That doesn’t mean AI is choosing the candidates, but it is giving recruiters otherwise inundated with applications with a short list off which to respond. 

One reason those recruiters are inundated with applications in the first place? Job seekers are also heavily leveraging AI and automation tools to secure the best chance of obtaining some kind of placement. 

Early in 2023, not long after ChatGPT was first released, a survey by Resume Builder showed that about half of job seekers were using the technology to help write their resumes or cover letters. Presumably it’s much higher now as the service gets more adoptions and as a cottage industry of software companies — Teal and EarnBetter among many others, as well as popular job sites like ZipRecruiter and LinkedIn — has emerged to formalize the process of using AI in the job search. That includes finding the best job openings for particular candidates, optimizing their applications, and suggesting follow-communications.

LinkedIn says that of the premium users who’ve had access to the AI experience so far, 90% have said it was useful and 70% have used AI-suggestions for writing and profile recommendations. It’s currently rolling out these experiences more broadly.

The number of job applications per job seeker are up 6% year over year, according to data provided by LinkedIn, though it’s impossible to tell how much of that is due to AI versus, say, a weaker job market. 

AI is certainly enabling some people to “spray and pray” their applications, a version of “one-click apply” on steroids. Candidates are using such technology to send out hundreds of applications, where they in the past would have only submitted dozens. Business Insider reporter Aki Ito recently employed AI-enabled job search bots to send out previously impossible levels of applications on her behalf. These tools sent out the wrong cover letter and made up lies about languages she spoke, but still she got a few calls back. 

On the recruitment side, such behavior is leading to increased ghosting, as would-be job seekers apply to positions they have no intention of really taking. In turn, recruiters are having to be more stringent in how they’re winnowing down the applicants they get.

Bot, meet bot.

An already infuriating process, but now faster, and with more spam!

It’s important to remember that the process of finding a job has long been bad.

Fuller, the Harvard Business School professor, said hiring practices have always privileged people with the same experience as the job they wanted. In other words, if you already were a tech project manager you’re more likely to be hired as a tech project manager, creating an infuriating cycle. That always excluded those who worked in adjacent industries, those without degrees or those who only had germane life experience, say someone who was in the Air Force who maintained jet aircraft applying to work at a repair depot for trucks.

“Usually what companies are doing is automating the decision rules that they have in their head or they followed back in the day when they were opening mailed-in resumes with letter openers,” Fuller told Sherwood. “A lot of the problems are baked in before anyone even starts using AI.”

By supercharging the process with AI, of course, there’s the potential for a lot more of the same. AI hiring boosters hope it can mean something different and better.

Fuller suggested using AI to pinpoint other experiences and skills that people have that might make them good — if not traditional — fits for certain jobs. That would give employers a larger pool of candidates than those with a specific degree or the exact job experience. 

In turn, AI tools can show candidates what they’re missing compared with other people applying for the role, so they can try and gain that experience, apply more selectively or make the case as to why they already fit.

One of the primary ways iCIMS is using AI is to help match candidates and companies — and vice versa — in a more expansive way than they might have on their own. 

If a job seeker uploads their resume to one of the many company websites that use the software, it directs them to appropriate jobs based not only on what they have listed explicitly in their resume but also on skills it infers they have but might not have noted.

“Because we have looked at hundreds of thousands if not millions of candidates with similar backgrounds, similar experiences in similar industries with similar titles, we know that if you tell us that you have three skills, you probably have the fourth and the fifth one as well,” Andreea Wade, VP of product strategy at iCIMS, said. “People are not great at building their CVs.” 

Wade sees AI as a marked improvement over older applicant tracking systems where users would simply use keywords to screen out candidates. 

“Applicant tracking systems that had zero intelligence in them were screening candidates out, because maybe some keywords did not exist there,” Wade said. “So we literally had technology that did not work.”

One of the major themes that arose from talking to job candidates using AI was the desire to “beat the applicant tracking system” — a foggy idea of how to fight the unseen forces, presumably AI, keeping them from getting interviews.

Jefferson Stovall, a public affairs executive, became so enamored with gen AI tools during his own job search that rather than getting a job he started a consultancy, Captain Tomorrow, to bring tech like gen AI to public sector communications. One prompt he’s fond of: feeding the gen AI the company and job description and then telling it to act like an applicant tracking system and surface the keywords most important for his resume and cover letter.

Brent Hametner, a marketing manager, uses gen AI to come up with custom demand generation strategy proposals for the companies he’s applying for — a request he says usually comes later in the interview process — to set himself apart from a lot of other applicants by going above and beyond what the initial applications require. He said he’s getting interviews from roughly 20% of his applications and is hoping to close in on a job soon.

A teacher we spoke with who is trying to transition out of education has been using ChatGPT and other AI software to apply to approximately 50 jobs in the past few months. He said it’s made writing cover letters drastically easier.

“AI just articulates it better,” he said, asking us not to use his name so it doesn’t affect his job search. He also credits AI with his “cognitive awakening,” teaching him to speak better and parse information more quickly. 

That said, it’s not gotten him very far on the job front, but that might have to do more with companies’ willingness to accept people without specific job experience rather than the strength of his cover letter.

“You could tell they're using AI to generate the thank you but no thank you letter.”

The end result: some parts of job process is going analog again.

While we wait for AI to go from more to better, AI fatigue is setting in. It’s left both employers and job seekers alike thirsty for a human touch.

As job emails and messages on job sites become increasingly cluttered and filled with spam and as more and more people pump resumes into what seems like a black hole, candidates and companies are going old-school. They’re reaching out to people through mutual contacts, meeting up in person, and asking for employee referrals. They’re networking.

“Candidates are going to try to find employers through warm contacts and employers that put in an effort beyond just posting a job posting online, who reach out to them to create a human connection,” Atta Tarki, author of Evidence-Based Recruiting, said.

He likened the situation to how some people, tired of dating apps, are trying their luck at bars and in speed dating.

“I know a lot of my young friends nowadays don't like using online dating communities because they feel like it just brings along a lot of flakiness,” Tarki said.

Employers are feeling the pain of AI too, making them seek out in-person or video conversations and find ways to make sure people are who — and as talented as — they say they are. 

“As a result of the reduced signal accuracy of a resume or a cover letter — the fact that any fool who can't read, write or spell can now write a perfect cover letter or a good enough cover letter — it makes it difficult for employers to weed out the weakest candidates,” ZipRecruiter Chief Economist Julia Pollak told Sherwood.

That’s meant more timed and recorded hiring assessments, as well as investment in products by companies like ZipRecruiter to make it easier and faster to have face to face connections

In short the rise of AI in the job search process has meant a corresponding rise in efforts to prove you’re not AI.