Key takeaways
- AI interviews bring speed, consistency, scale and flexible scheduling, especially for early high volume screening.
- The downsides include poor candidate experience, bias risk, strong people filtered out and candidates gaming the format.
- Some places regulate AI in hiring, such as Illinois with its Artificial Intelligence Video Interview Act, which requires notice, explanation and consent.
- Use AI to structure and collect answers, keep final decisions with trained people, and always offer a human alternative.
What we mean by AI interviews
"AI interview" covers a few different things, and its worth being clear, because the risks are different for each one.
- One way video interviews: the candidate records answers to set questions, with no live person on the other side. Some platforms only record. Others analyze the answers.
- AI screening bots: chat or voice assistants that ask screening questions, collect answers on availability, location, work authorization and skills, and sometimes ask short technical questions.
- Automated scoring: software that ranks or scores candidates based on their answers, resume or assessment results, and in some tools the speech and language in a recording.
Most employers use a mix. And many candidates don't know which kind they are facing, which is part of the problem.
The positive impact
Speed
A recruiter can only do so many phone screens in a day. A one way video or chat screen can run overnight for hundreds of applicants. For high volume roles, or a requisition with a tight start date, that speed has real value.
Consistency
Every candidate gets the same questions in the same order. Human screeners have good days and bad days, and their questions drift over time. A structured AI screen, if it is designed well, gives each person the same starting point.
Scale
Companies that hire across many locations or run large programs can screen at a scale that would otherwise need a much bigger recruiting team.
Scheduling and time zones
A candidate can complete a one way interview at ten at night after their current job. For consultants who are working full time, or offshore candidates in a different time zone, that flexibility is a real benefit. I have seen it help people who would never get a free weekday afternoon.
The negative impact
Candidate experience
Many candidates find talking to a camera with nobody there awkward and impersonal. Senior people especially. A senior architect with other options may simply skip your process. So the tool that saves you time can quietly shrink your pool of top talent.
Bias risk
AI learns from data, and hiring data has history in it. A model trained on past hires can repeat past patterns. Tools that judge speech, accent, facial expression or video quality raise extra concern, because those things have little to do with whether someone can do the job. Candidates with disabilities, non native English speakers and people with weak internet can be scored down unfairly.
Good people filtered out
Automated knockouts are blunt. A strong candidate answers one question in an unexpected way, or describes a skill with different words, and never reaches a human. The worst part is you never find out who you lost.
Candidates gaming the format
Candidates learn quickly. Some read scripted answers off a second screen, stuff keywords into every reply, or use tools that suggest answers in real time. So a high score can mean someone is good at the format rather than good at the job.
Legal and notice requirements
Some places regulate AI in hiring. Illinois has the Artificial Intelligence Video Interview Act. In general terms, an employer that uses AI to analyze video interviews for positions based in Illinois must tell applicants beforehand that AI may be used, explain how it works and what general types of characteristics it evaluates, and get their consent. The law also limits who the videos can be shared with and requires deletion when an applicant asks. Employers that rely only on AI analysis to decide who gets an in person interview also have demographic reporting duties to the state.
New York City has a law on automated employment decision tools that requires bias audits and notice to candidates for covered tools. Other states and countries have their own rules in place or in progress. This is not legal advice, and these rules change, so check with your counsel for each place where you hire.
How to use AI interviews fairly
- Tell candidates what is happening. Say clearly if a tool records, analyzes or scores them, and what it looks at. Do this even where the law does not require it.
- Use AI for structure, and keep the decision human. Let the tool collect and organize answers. Let a trained person make any decision to reject.
- Score job related content only. Avoid tools that score facial expressions, tone of voice or background. Focus on what the person said and whether it answers the question.
- Offer an alternative. Give candidates a way to request a live interview or an accommodation. Some will need it, and some of your best candidates will simply prefer it.
- Check outcomes regularly. Where you lawfully can, look at who passes and who fails by group. If one group fails more often, investigate the tool.
- Keep the screen short. Three to five relevant questions is enough. Long one way interviews lose candidates fast.
- Ask the tool vendor hard questions. How was the model tested for bias? What data is kept, and for how long? Who can see the videos?
Advice for candidates facing an AI interview
If you are a candidate, don't panic about the format. Treat it like a real interview. Good light, a quiet room, camera at eye level. Answer the question that was asked, with a short real example from your work. Speak naturally and don't try to stuff keywords, in most processes a person reviews the shortlist at some stage anyway.
And if you need an accommodation, ask for it. Many employers have a process for that. If you are not sure whether AI is being used, it is fine to ask the recruiter directly. A good recruiter will tell you.
Where AI interviews fit best
From the staffing side, I think AI interviews work best for early, high volume screening with clear requirements. They work poorly for senior, niche or consulting roles, where a twenty minute real conversation with a recruiter who knows the domain tells you far more. For a solution architect or a healthcare revenue cycle manager, I would not let a bot decide who moves forward. For those roles a human screen by someone who knows the work is still the best filter. You can see how we approach screening on our IT staffing services page.
Frequently asked questions
Is it legal to use AI in job interviews?
Generally yes, but some places add rules. Illinois requires notice, an explanation and consent before AI analyzes video interviews, and New York City requires bias audits and candidate notice for covered automated hiring tools.
Can AI interviews be biased?
Yes. AI tools can repeat patterns from past hiring data, and tools that judge speech, accent or facial expression can disadvantage some groups, so employers should check outcomes and focus scoring on job related content.
How do I prepare for an AI video interview?
Treat it like a live interview: quiet room, good light and camera at eye level. Answer each question directly with a short real example and speak naturally instead of stuffing keywords.
Want screening that real people stand behind?
Our recruiters screen every candidate personally before submittal, with AI used to support the process and never to replace judgement.
Israr Kazmi
Israr Kazmi is the founder of iCareTechno, the IT staffing, recruitment and consulting division of iCareBilling LLC in Chicago. He writes about hiring, vendor relationships and careers in IT.
More about the author