When Job Interviews Become Bots Talking to Bots

Christopher has applied to roughly 700 jobs over the past six months. His work as a government contractor dried up during the DOGE-era cuts, and like millions of other job seekers, he's learned to accept silence as the default response. Most applications vanish into the void. But five of them didn't just vanish, they led him somewhere stranger: a series of conversations with an AI recruiter named Riley, built for the IT firm Everforth Apex Systems, that ultimately convinced him the entire hiring process had become theater.

The first call felt like a small victory. It was June, and Christopher, who asked WIRED to use only his first name while his job search continues, was grateful for any chance to make his case, even to a machine. Riley asked about his work authorization and background, then promised a human recruiter would follow up if he qualified. Nobody did. A little over a week later, Riley called again about "another opportunity." Then again. And again.

By the fifth call, something in Christopher snapped, but not into anger exactly. Into a kind of dark curiosity. If the company was comfortable outsourcing its side of the conversation to software, he figured, why should he keep doing the human labor of pretending to be present and engaged? So he fed a few sentences of his background into ChatGPT Voice, told it to answer as "Christopher," and let Riley call. Then he stepped back and watched two machines talk to each other for ten minutes.

The exchange was almost comically polite. Riley told ChatGPT it was "great talking to you." ChatGPT replied that it appreciated "how clear and straightforward the process was." At one point the two bots got stuck in a loop trying to pin down a start date, circling back again and again to "standard onboarding and background checks" without ever landing on an answer. Riley made the same promise it had made four times before: a recruiter would reach out if Christopher's application met the company's criteria. No recruiter ever did.

Christopher has a term for what happened. He calls it a "slop flywheel," a synthetic voice feeding synthetic answers to another synthetic voice, generating a transcript that serves no one. "This is one synthetic persona giving slop data to another synthetic persona," he told WIRED. "And all of the data is going, where? Nowhere." He described the experiment as equal parts mischievous fun and genuine frustration, and ultimately, a signal that he had no real interest in working for a company that treated hiring this way. Everforth Apex Systems did not respond to WIRED's request for comment.

Why Recruiters and Applicants Both Turned to Bots

AI has crept into nearly every stage of the job hunt, but voice interviews took longer to catch on than resume screening or application filtering. Voice AI is notoriously hard to get right. Agents stumble over accents, hesitations, and the small verbal cues, like knowing not to interrupt, that come naturally to a human interviewer. Ophir Samson, who heads voice AI at the recruitment platform Greenhouse, describes even simple conversational courtesy as a "very, very difficult engineering problem" to solve in code.

Yet the pressure on the other side of the hiring desk has been enormous. Recruiters, buried under a flood of applications, have leaned into AI screening tools out of sheer necessity. Greenhouse reports that 63 percent of job seekers have now encountered an AI interview at some point in their search. That statistic alone tells you how normalized this has become. What used to be a novelty is now simply part of the process, an early filter standing between applicants and any human contact at all.

Naturally, job seekers adapted. Using AI to help draft cover letters or rehearse answers has become common enough that recruitment startups like Ribbon now market tools specifically designed to catch "overly scripted, AI-assisted, or coached" responses. In other words, the same industry that automated the interview is now racing to detect when candidates fight back with automation of their own. Christopher's experiment sits right at that fault line.

Mark Monaghan, vice president of organizational development at the call center company IQor, sees the collision as almost inevitable rather than scandalous. He calls bot-on-bot interviews the "next logical stage" of a hiring process that has been drifting toward automation on both sides for years. That doesn't mean he's comfortable with where it's landed. But given how Everforth Apex handled Christopher's applications, Monaghan believes the workaround was more than fair. "If you're going to send a bot to me," he says, "I'll send a bot to you."

Christopher wasn't finished testing the system. After his fifth dead-end interview, he wondered if the failure was somehow his own, if perhaps a more qualified candidate would get a different response. So he invented one: a fictional applicant named Don Dickner, whose resume he built entirely from the language of an open job posting, packed with exactly the qualifications the listing wanted.

Riley called almost immediately. This time the conversation ran 23 minutes, with ChatGPT posing as Dickner and fielding questions about "sustainable operational improvement," protecting "customer experience under volume pressure," and never letting "tribal knowledge drift." ChatGPT had a smooth answer and a fabricated anecdote for every prompt. As the call wound down, it even asked if Riley had "a few more questions I'd like answered." Riley's reply was polite and familiar: the application would be reviewed, and a recruiter would follow up if Dickner met the company's criteria. Christopher never heard from Riley again.

Why It Matters

What Christopher's experiment reveals isn't just one company's broken process. It's a preview of what happens when both sides of a critical human decision, whether someone gets hired, whether someone can pay rent, get outsourced to software that was never designed to actually decide anything. The interviews generated transcripts, timestamps, and polite exchanges, but no real evaluation ever took place. For the hundreds of thousands of job seekers now encountering AI interviews, the unsettling question isn't whether a bot can convincingly imitate a conversation. It's whether anyone, human or otherwise, is actually listening on the other end.