Here’s some recruiter math nobody enjoys. Two hundred applicants for one customer-support job. Twenty minutes per phone screen. That’s 66-plus hours of calls, and nobody has met a hiring manager yet. Add the reschedules, the no-shows, and the applicant two time zones away who can only talk at 7 a.m.
Most of those calls exist to answer three questions. Can this person do the job on paper? Can they explain themselves? Are they still interested?
One-way video interviews skip the calendar juggling. Candidates record answers to set questions when it suits them, and recruiters watch later, usually with coffee and the playback at 1.25x. Tools built for this, like Hireflix, revolve around that handoff: the candidate gets a link and records, and the hiring team reviews and shares the answers whenever it has time.
There’s another side to this format, and it’s the more interesting one. Recorded interviews are where AI hiring tools first went mainstream. They’re also where those tools got into the most trouble. So anyone adopting the format now has a second question to answer: how much of the judging should software do?
What Problem Do One-Way Interviews Actually Solve?
Not the conversation. The scheduling.
For one vacancy, honestly, a few phone calls are fine. Fifteen similar roles across three cities is a different story. Every new applicant adds more back-and-forth emails, while sending a one-way invite takes the same effort for 10 people as for 500.
Candidates feel the difference too. “Record a quick interview tonight” gets done. “Let’s find a half-hour next week” gets put off, and a strong candidate can sign somewhere else in the meantime.
How Does a One-Way Interview Work?
The recruiter writes the questions, picks a time limit for each answer, and sends the link. The candidate sees a question, records, and moves on. That’s about it.
Where it goes wrong is setup. Make someone download an app or create an account to answer four questions, and a fair number won’t bother. They’ll close the tab and apply to the next job in their feed. The good tools make recording close to effortless.
Why Does Consistency Matter for Fair Comparison?
Think about the last time you ran six interviews in a day. By the fourth one, you were probably asking questions a bit differently, following whatever tangent came up, maybe skipping a question because you were running late. Nobody does this on purpose. It just happens.
Recorded screens mostly stop that. Same questions, same time limits, for everyone. A hiring manager can watch five people answer the identical prompt one after another and actually compare them. Industrial psychologists have said for decades that structured interviews, where every candidate gets the same questions, predict job performance better than loose conversations.
That consistency is the whole selling point. It’s also what makes automation so tempting.
Where Does AI Fit Into One-Way Interviews?
Once the answers sit in a database, someone always asks: can’t software score these?
HireVue tried. The best-known vendor in the space once analyzed candidates’ facial expressions as part of its assessments. Then came criticism from researchers, and in 2019 an FTC complaint from the Electronic Privacy Information Center. In January 2021, HireVue said it had already dropped the facial analysis, arguing that language analysis had made the visual data mostly redundant. Emotion scientists had been saying all along that a face, on its own, doesn’t tell you much about character or ability.
None of this means AI has no role. It does. But “help a human review faster” and “decide who moves on” are very different jobs.
| What the AI does | How risky | The short version |
| Transcripts and captions | Low | Faster review and better accessibility. A person still judges. |
| Searchable answers and summaries | Low to medium | Handy, but summaries flatten nuance. Spot-check them. |
| Flagging broken recordings | Low | Nobody should lose a job over bad Wi-Fi. |
| Scoring or ranking candidates | High | This is what the hiring laws below target. |
| Reading faces or emotions | Don’t | Shaky science, and banned in EU workplaces. |
| Auto-rejecting people | High | It cuts the human out exactly where it matters most. |
Candidates already notice. Graduates describe job hunts where résumés get filtered and scored before a human ever looks. A recorded interview graded by a model feels like more of the same. One watched by a person feels like a fair shot.
What Do AI Hiring Laws Require?
If AI is scoring anywhere in your setup, some of this probably applies to you already.
Illinois got there first. Its Artificial Intelligence Video Interview Act, in effect since 2020, says that if AI analyzes a candidate’s video, the employer must tell them, explain how it works, and get consent. It also limits who sees the footage and requires deletion on request.
New York City started enforcing Local Law 144 in July 2023. Automated hiring tools need an independent bias audit every year, a published summary of the results, and advance notice to candidates.
Colorado has struggled with this one. Its AI Act has been delayed and rewritten several times. The current version aims for 1 January 2027 and leans on notice and transparency instead of audits, and a court has paused enforcement for now. Check the status again before you rely on it.
The EU treats AI used in recruitment as high-risk under the AI Act. After the Digital Omnibus delay, most of those obligations arrive on 2 December 2027. The ban on emotion recognition at work already took effect in February 2025.
Put simply, recording a video is rarely the legal problem. Letting an algorithm decide who goes through is.
How Is AI Changing the Candidate Side?
Recruiters aren’t the only ones with AI. Gartner found that 39% of candidates used it during their applications, mostly for résumés, cover letters and test answers. Only 26% trusted AI to judge them fairly. So both sides are using a tool that neither side really trusts.
It gets worse. In another Gartner survey, 6% of candidates admitted to interview fraud: posing as someone else, or having someone else pose as them. Gartner expects one in four candidate profiles worldwide to be fake by 2028. Voice-security company Pindrop caught an applicant for a senior engineering role, “Ivan,” using deepfake software to change his face during a video interview. Hiring has quietly become part of the same generative-AI security fight that companies are dealing with everywhere else.
That undermines the consistency argument. If one person reads a polished AI script off a second monitor and another answers honestly, identical questions no longer produce a fair comparison.
The defenses aren’t complicated. Ask about things only the candidate could know, like a specific project or a specific bad day. Keep time limits short enough that reading a script shows. Check identity before an offer, not only at the application stage. And make sure everyone who passes the recording stage talks to a real person.
How Do You Write Better Screening Questions?
A weak question set wastes everything above, so it’s worth getting right.
Make the questions behavioral and specific to the role. “Tell us about a customer who stayed angry after you’d fixed the problem” tells you far more than “What are your strengths?” Three to five questions is enough for a first screen. Go past that and both the candidates and whoever reviews forty videos will be worn out.
Time limits should allow a real answer without letting it ramble. Two minutes works for most prompts. Anything that needs a follow-up question belongs in the live round. And after each hiring cycle, cut the prompts that didn’t separate strong candidates from weak ones. There are usually one or two.
Where Should One-Way Interviews Sit in the Hiring Process?
Up front. As a filter, not a verdict.
The teams that get the most from recorded screens use them to cut a pile of 200 down to 15, quickly and fairly, and then switch to live conversations. That’s where you can push on an interesting answer, see how someone handles back-and-forth, and confirm the person on screen is who they claim to be.
FAQs
Q. Are one-way video interviews fair to candidates?
Often fairer than improvised phone screens, since everyone gets the same questions. Fairness drops when unaudited AI scoring makes the call, or when nobody gets a live conversation before a decision.
Q. Do these platforms use AI to score candidates?
Some do, and plenty don’t. Ask the vendor directly, because the answer decides which hiring laws you need to follow.
Q. Can candidates cheat?
Some will try, with AI-written scripts or occasionally deepfakes. Experience-based questions, tight time limits, and a live interview before any offer make it much harder.
The Bottom Line
One-way interviews fix a boring but real problem: too many calls and not enough hours. Used well, they also make early screening fairer than a stack of improvised phone calls. Problems start when teams let the recordings make the decision. Let software handle the scheduling, the transcripts, and the logistics. Keep people in charge of deciding who moves forward.
Related: How to Hire Remote Tech Talent in Latin America: 8 Steps
