The job interview is drifting beyond office hours. As employers put automated video and voice systems at the entrance to their hiring funnels, candidates are recording answers late at night—sometimes because flexibility helps, and sometimes because a human conversation has been replaced by an obligation that can be completed at any hour.
WIRED describes communications specialist Tim Millard preparing his living room for a recorded interview at 9:30 pm. Questions appeared on his laptop, followed by about 30 seconds to think and two minutes to answer. He later received an automated rejection.
His experience captures the ambiguity of AI-assisted recruitment. An asynchronous interview can open a door for someone who cannot leave work or childcare during the day. It can also make an already impersonal job search feel even more remote.
Convenience is measurable, but so is candidate discomfort
Recruitment software provider Ribbon says 24 percent of interviews conducted through its AI system occur between 10 pm and 2 am in the candidate’s local time. Among its manufacturing customers, the figure rises to 35 percent. Greenhouse reports that roughly 15 to 20 percent of candidates using its voice agent schedule an interview at night.
Those company figures suggest that availability is not a trivial feature. Parents, hourly workers and people interviewing while employed may have few uninterrupted periods during conventional business hours. A system that waits for them can remove a real scheduling barrier.
Yet flexibility does not automatically equal preference. Millard wanted to finish an uncomfortable requirement, not celebrate a new freedom. Greenhouse survey results cited by WIRED show a wider trust problem: nearly two-thirds of applicants said they had encountered an AI interviewer, while 38 percent reported leaving a hiring process rather than being interviewed by one. Another 12 percent said they would withdraw if it became mandatory.
Scoring turns transparency into the central issue
Automated interviews are generally evaluated against a role-specific rubric. A technical candidate might be assessed on certifications or job knowledge, while a sales applicant could receive an enthusiasm score. The system analyzes the recording, assigns a result and sends it to a recruiter alongside the footage.
What candidates often do not know is how much authority that score carries. They may be unsure whether a missing phrase can remove them, whether a person will watch the recording, or how voice and video data will be retained. WIRED notes that many submitted interviews may never receive human viewing.
Vendors offer a more optimistic interpretation. Greenhouse executive Ophir Samson described the interview as an addition to a résumé that can identify promising candidates buried in a flood of applications. In this framing, the alternative is not necessarily an immediate human interview; it may be no response at all.
That remains a vendor argument, and outcomes depend on employer practice. A screening tool that recommends candidates for review is materially different from one that silently becomes a rejection gate.
A human next round does not solve an inhuman first one
Employers considering these systems need to explain the process before asking candidates to participate. Applicants should know what is measured, whether an automated score can eliminate them, who can access the recording and when a person will become involved. An alternative assessment route would also help candidates who cannot or do not want to submit video to an AI system.
Recruiters must examine whether late-night usage reflects welcome flexibility or pressure from people who feel unable to refuse. Completion rates and scheduling statistics reveal behavior, but not whether candidates regarded the experience as fair.
Millard’s story ends with the emotional cost of a prolonged search rather than a technical verdict on one platform. After losing his job, he was still looking a year later and turned the experience into a one-person show. His recorded interview became another encounter with a process that felt like speaking into a void.
AI can make the first hiring step available around the clock. The harder task is ensuring that availability does not become a convenient way to remove accountability from the people making employment decisions.
