Job applicants are winning the AI arms race against recruiters

Many harried recruiters would have greeted the release of ChatGPT with glee. At last, a tool that could ease the burden of drafting job descriptions, scheduling interviews and rejecting candidates. Like happy beachgoers who cannot see an approaching tsunami, they failed to predict what came next. A giant wave of AI-generated applications has washed over them.

The number of job applications an average candidate sends has risen by 239% since ChatGPT’s release in 2022, according to global data collected by Greenhouse, a provider of applicant-tracking software. Paid services like LazyApply and aiApply let candidates submit applications while they sleep, tailoring résumés and cover letters to a tee. AI has even made it easier for spies and fraudsters to infiltrate companies. Last month Amazon blocked 1,800 applications from North Koreans applying for remote IT jobs. Gartner, a consultancy, predicts that by 2028 as many as one in four candidate profiles could be fake.

Inundated recruiters are taking various steps in response. Some are politely asking candidates to refrain from using chatbots. Anthropic, an AI lab, asks candidates not to submit entirely AI-generated cover letters. So does Mastercard, a payments provider. Others are capping the number of applications a person can submit. OpenAI, another AI lab, limits candidates to a maximum of five over a six-month period.

Companies are also speeding up their use of AI to help winnow the growing pile of résumés and cover letters. Two-thirds of recruiters plan to increase their use of the technology to conduct screening calls, according to data from LinkedIn, a social network for professionals. Many also use AI to sift through applications and flag those that match the criteria for a job, though firms such as KPMG, a professional-services giant, are keen to emphasise that humans make final decisions. AI models can even recommend applicants for one role as candidates for another, says Alicia Pittman, who directs hiring for BCG, a consultancy. Still, the time it takes for companies to fill vacancies has declined only slightly since 2021, according to Ashby, a recruiting-software company.

When it comes to adopting AI, recruiters face a structural disadvantage compared with job applicants, reckons Robert Newry, co-founder of Arctic Shores, a psychometric testing firm. Jobseekers need not worry whether an AI tool will run afoul of anti-discrimination or data-protection laws. They do not need to check with their boss or IT department before adopting the latest technology. “I say to my clients: ‘In the arms race between you and the candidate, you will lose,’” says Mr Newry.

Eventually the rise of AI may change hiring more fundamentally. Companies could rely more on tasks that cannot be pasted into chatbots, such as visual puzzles, says Mr Newry. They might also rely more on finding candidates before they even think to apply. Juicebox, a recruitment startup, offers a service called “PeopleGPT” that automates the process of hunting for possible hires across the web. It says that many of its clients have reduced the time it takes to find candidates by half. LinkedIn has rolled out a “Hiring Assistant” for recruiters that trawls the site for suitable candidates.

Companies may one day abandon the job application entirely, to the delight of recruiters and candidates alike. Daniel Chait, boss of Greenhouse, reckons that both sides will eventually use AI agents to talk to one another and determine whether they might be a match. “We’re at the point today where it’s automating tasks,” Mr Chait says. “I’m eager to get to the part where we can think about why we have job posts at all.” Until then, recruiters will have to wade through a sea of em-dashes and eerily polished résumés as they search for the perfect hire. ■