Job interviews are no longer confined to office hours. As automated video interviews and AI voice agents move deeper into hiring, candidates are increasingly finding themselves answering screening questions late at night, sometimes after work, childcare or other obligations have ended for the day.
The change is practical for some applicants. It is also unsettling. For people already worn down by a long job hunt, talking to software instead of a person can make the process feel more distant, less transparent and harder to trust.
A hiring step with no human in the room
Tim Millard, an Atlanta-based communications and marketing expert, encountered that reality in January while interviewing for a media relations job. He had been searching for work for six months when an email arrived about an interview. He expected a conversation with a person.
Instead, he was asked to record himself answering questions on camera. The schedule was open-ended, so he prepared his living room that evening, set up his laptop and completed the interview at night. As he put it, “What else was I going to do at 9:30 at night?”
The questions themselves matched the role. Millard was asked how he would pitch a story to the media and how he would build relationships with local journalists. The strange part was the format. A question appeared on screen, he had roughly half a minute to think, and then a timed recording window began.
“Once the 30 seconds was up, ‘ding!’ Now you have two minutes to record your answer,” Millard said. “It felt so out of this world.”
In March, he received an automated rejection email. His experience sits inside a broader shift: late-night interviews are becoming more common as hiring platforms make screening available outside the normal workday.
Why interviews are moving later
Arsham Ghahramani, who cofounded and launched Ribbon, a company that makes voice-AI recruitment software, said the late-night pattern appeared quickly after launch. According to Ribbon data from the more than 500 companies using the service, 24 percent of its AI interviews happen between 10 pm and 2 am local time.
Among Ribbon’s manufacturing clients, the number is higher: 35 percent. That suggests the appeal may be especially strong for workers whose schedules make ordinary interview windows hard to manage.
Ghahramani describes the late-night option as useful for people with limited scheduling flexibility. Parents may not have 30 minutes free until 10 pm. Hourly workers may not be able to step away during the day. Others may be working current jobs in noisy settings such as a factory floor or kitchen.
Hiring platform Greenhouse is seeing a similar pattern. The company says roughly 15 to 20 percent of candidates interviewed by its voice agent now schedule an interview at night.
In that sense, AI job interviews can solve a real logistical problem. They let candidates complete a screening step when a recruiter is not available and when the applicant finally has privacy, quiet and time.
Flexibility comes with distrust
The convenience does not erase candidate skepticism. Ophir Samson, who became head of Greenhouse’s voice AI division when the company acquired his startup in May, told WIRED that he does not think artificial intelligence should replace human interviews.
That distinction matters because applicants are wary of the technology. A Greenhouse survey in May found that nearly two-thirds of job applicants have been interviewed by an AI agent, up 13 points over six months.
The same survey showed resistance. Thirty-eight percent of American candidates said they had withdrawn from a hiring process rather than be interviewed by an AI. Another 12 percent said they would leave the process if an AI interview were required.
One major concern is transparency. Candidates may not know what will happen to their video or audio, who will review it, or whether software can screen them out because their answers fail to match expected language.
The source article describes a typical AI or automated video interview as rubric-based, though the details vary by employer and service. The rubric changes by job. A welder might be asked about technical skills or certifications in a Ribbon interview, while a sales role might include an “enthusiasm” score.
AI then analyzes the interview footage, measures the candidate against the rubric and assigns a score. That score is sent to the recruiter along with the recording. In practice, many of those interviews may never be watched by a person.
The promise and the emotional cost
Greenhouse frames its voice-AI interview as a kind of résumé supplement. Samson’s view is that the tool can help identify strong candidates who might otherwise disappear into a flood of applications. In his framing, the issue is not simply human interviewer versus AI interviewer. It is whether a candidate is ignored or gets a better chance of reaching a human in a later round.
He described the job market in blunt terms: “It’s a shitty time to be a candidate.” He added that résumés can feel like they vanish, while applicants compete with thousands of others who are also using AI to write their résumés.
Millard’s experience shows the other side of that logic. For him, the late-night interview was not an exciting new convenience. It was another required step in a difficult search. A year after losing his job, he was still looking, and the strain led him to write a one-man show, After Careful Consideration, which debuted at the Atlanta Fringe Festival in May.
When he completed the January interview, Millard said he simply wanted to get “this godawful thing done.” His broader description of the market was even sharper: “In today's job market,” he said, “it feels like if you are going to scream, you're going to scream into the void. And nobody is going to hear you.”
That is the tension at the center of late-night AI interviewing. The technology can open interview windows that would not otherwise exist. But if candidates do not understand how they are being judged, and if recordings often go unwatched, the process can also make hiring feel more automated at the very moment applicants are trying to be seen.