A government contractor who applied to roughly 700 jobs over six months turned to an unusual tactic after being ignored by an AI recruiter five times: he let ChatGPT conduct his sixth interview with the same automated system. The experiment, detailed in a report by WIRED, highlights how artificial intelligence has infiltrated both sides of the hiring process—and what happens when two bots meet with no humans involved. Christopher, a contractor whose work dried up during recent federal budget cuts, describes the resulting exchange as a "slop flywheel" where synthetic personas swap meaningless data that goes nowhere.
Christopher's journey began in June when Riley, an AI recruiter for IT staffing firm Everforth Apex Systems, first contacted him by text. Over the following months, Riley interviewed him five separate times for different positions, asking about work authorization and professional background before promising a human would follow up if he qualified. No recruiter ever reached out, nor did he receive automated rejection messages. By his sixth opportunity with the same company, Christopher fed a few sentences of background information into ChatGPT Voice and instructed it to impersonate him during the call. The two AI systems conversed for 10 minutes, with Riley telling ChatGPT it was "great talking to you" and the chatbot praising how "clear and straightforward the process was." They cycled through a circular discussion about start dates and background checks before Riley delivered the same empty promise she'd made four times before. Christopher also created a fictional candidate named Don Dickner with a résumé stuffed with qualifications lifted directly from a job posting; ChatGPT, posing as Dickner, spoke with Riley for 23 minutes about operational improvement and customer experience, fielding every question with manufactured anecdotes. Neither the real Christopher nor the fake Dickner ever heard from a human recruiter.
Voice AI has struggled to gain traction in recruitment interviews because agents stumble over accents or filler words, and tasks that come naturally to humans—like not interrupting a candidate—become complex engineering challenges, according to Ophir Samson, who leads voice AI at recruitment platform Greenhouse. Still, recruiters drowning in application volume have increasingly adopted these tools: 63 percent of job seekers report encountering an AI interview, per Greenhouse data. Candidates have responded in kind, using AI assistance so frequently that startups like Ribbon now promise to detect "overly scripted, AI-assisted, or coached" responses identical to those Christopher deployed.
Some industry observers view bot-on-bot interviews not as deception but as an inevitable development. Mark Monaghan, vice president of organizational development at call center operator IQor, calls it the "next logical stage" and justified Christopher's AI use with a simple principle: "If you're going to send a bot to me, I'll send a bot to you." The dynamic creates a feedback loop where neither side invests genuine human attention, producing what Christopher labels synthetic personas exchanging worthless information. After his ChatGPT experiment yielded the same non-response as his own efforts, he dismissed any interest in working with the company that had wasted his time across six futile interactions. The rise of AI on both ends of recruitment may force employers to reconsider whether automated screening actually identifies talent or simply trains candidates to game systems that were meant to save time. When machines interview machines and no one follows up, the entire process risks collapsing into theatre where neither party expects a real outcome.