AI Interviews: What Candidates Should Know About Automated Interviews

Published · by JOBB GROUP FZE

An AI interview is any interview stage in which software, rather than a person, asks the questions, records your answers or helps score them. For candidates, the important point is that the software is usually one input into a human decision, not the final word, and in some places the law gives you specific rights to be told about it, to understand it and to ask for a person to look again. This article explains how these systems tend to work, what the rules say in the jurisdictions that have addressed them, and how to prepare.

What “AI interview” can mean

Employers and vendors use the phrase for several different things. Knowing which one you face changes how you prepare.

Format What happens What is likely to be assessed
Chatbot screening A text chat asks about eligibility, availability, location and basic experience Whether you meet stated requirements; some ask short competency questions
One-way video with automated scoring You record answers to on-screen questions within a time limit Usually a transcript of what you said, matched against criteria the employer has set
AI voice or avatar interviewer A synthetic interviewer asks questions and may ask follow-ups based on your answers Content and structure of answers; some tools summarise for a human recruiter
Game-based or task assessments Short puzzles, simulations or work samples Problem-solving, attention or job-specific skills, scored by the system

A recorded video interview is not necessarily an AI interview. Many employers simply have recruiters watch the recordings. If you are unsure, ask (see the questions later in this article).

How the scoring typically works

Vendors rarely publish full technical detail, so treat any description, including this one, as general. Most systems follow a similar pipeline:

  1. Capture. Your text, audio or video is recorded.
  2. Conversion. Speech is transcribed into text. Poor audio, heavy background noise or very fast speech can produce transcription errors.
  3. Comparison. The content is compared with what the employer has defined as a strong answer: specific competencies, examples of particular behaviours, or job-related knowledge.
  4. Output. The system produces a score, a ranking or a summary for a recruiter.
  5. Human decision. A recruiter or hiring manager reviews the output and decides who moves forward. How much weight they give it varies widely.

Some tools have claimed to read emotions or personality from facial expressions or tone of voice. In the European Union, the AI Act has prohibited AI systems that infer emotions in the workplace since 2 February 2025 (Article 5(1)(f)), and the European Commission’s guidelines, as reported by the Future of Privacy Forum, clarify that hiring processes fall within the workplace context for that prohibition. There are narrow exceptions for medical and safety reasons.

What the rules say

There is no single global rule on AI in hiring. The examples below are the clearest ones, as of November 2026. This is general information, not legal advice; check with the relevant regulator or a qualified adviser for your situation.

New York City: Local Law 144

New York City regulates “automated employment decision tools” used for jobs located in the city, including remote jobs linked to an office there. According to the city’s Department of Consumer and Worker Protection (DCWP), the law took effect on 1 January 2023 and DCWP began enforcing it on 5 July 2023. Key points for candidates:

  • The tool must have had an independent bias audit within the past year, and the employer or agency must post a summary of the results on the employment section of its website.
  • Candidates who live in New York City must be told that a tool is being used and which job qualifications and characteristics it assesses, at least 10 business days before it is used.
  • The notice must include instructions for requesting a reasonable accommodation under other laws.
  • Complaints about audit or notice failures can be made through 311 or the DCWP website; discrimination claims go to the NYC Commission on Human Rights.

Illinois: Artificial Intelligence Video Interview Act

For positions based in Illinois, an employer that uses AI to analyse recorded video interviews must, before the interview, tell the applicant that AI may be used, explain how it works and the general types of characteristics it uses, and obtain the applicant’s consent (820 ILCS 42/5). The section states that an employer may not use AI to evaluate applicants who have not consented.

European Union: AI Act

Besides the emotion-recognition prohibition above, the AI Act (Regulation (EU) 2024/1689) classes AI systems intended for the recruitment or selection of people as high-risk (Annex III, point 4(a)). Employers using such systems will have to assign human oversight to trained staff (Article 26(2)) and tell people that a high-risk system is being used in decisions about them (Article 26(11)).

The timing matters. These high-risk rules were due to apply from 2 August 2026, but the Digital Omnibus on AI (Regulation (EU) 2026/1744), in force since 27 July 2026, moved the date for Annex III systems to 2 December 2027, according to the European Commission’s implementation timeline. Until then, the emotion-recognition ban and the EU’s general data protection rules are the main protections for candidates.

United Kingdom: data protection

The UK Information Commissioner’s Office (ICO) gives “a recruitment aptitude test using pre-programmed algorithms and criteria” as an example of automated decision-making. Its guidance for the public says people have the right not to be subject to a decision based solely on automated processing where it significantly affects them, and that organisations should offer simple ways to get an explanation, request human intervention, express a view and challenge the decision. The ICO notes this guidance is under review following the Data (Use and Access) Act 2025, which changes the UK rules on automated decisions, so check the current position before relying on it.

In November 2024 the ICO published findings from audits of AI recruitment tool providers. It found some tools allowed recruiters to filter out candidates with certain protected characteristics, some inferred gender and ethnicity from a candidate’s name, and some collected more personal information than necessary and kept it indefinitely. The ICO made recommendations to providers and employers, including clear privacy information for candidates.

Everywhere else

If you are applying elsewhere, including in the UAE, read the employer’s privacy notice and the assessment instructions. They should explain what data is collected and how it is used. If they do not, ask.

Questions you can ask the employer

Asking these politely does not count against you with a reasonable employer. Send them to the recruiter, not through the assessment platform.

  • Is any part of this stage scored automatically, or will a person review my responses?
  • What skills or competencies is the assessment designed to measure?
  • Is video or voice analysed, or only the content of what I say?
  • How long will my recording and data be kept, and who can access it?
  • Can I request an adjustment, such as extra time, captions or an alternative format?
  • If I am not progressed, can a person review the decision?

How to prepare

Answer the question that is asked

Because most systems assess the content of your answers against defined criteria, relevance matters more than polish. Read or listen to the full question, identify what it is really asking for (a skill, an example, a judgement), and answer that directly in your first sentence.

Use clear structure

A structured answer such as situation, action and result is easier for both software and humans to match to a competency. Name the skill in plain words: “This was mainly a problem of prioritisation.”

Illustrative opening: “The clearest example of handling a difficult customer was in my last role at a hotel front desk. A guest’s booking had been cancelled in error during a fully booked weekend. Here is what I did…”

Use the job’s own language, honestly

If the advert asks for “stakeholder management”, use that phrase when it genuinely describes what you did. Do not repeat keywords without substance. Recruiters reviewing a transcript can see when a term is used without an example, and a structured interview later will test it.

Make yourself easy to transcribe

  • Use a headset or a quiet room. Background noise can cause transcription errors.
  • Speak at a steady pace and finish sentences.
  • Spell out or explain acronyms the first time: “an HACCP plan, that is a food safety plan”.
  • Check your microphone with a test recording before you start.

Request adjustments early

If a disability, a speech difference or a technical limitation could affect your performance in an automated format, ask for an adjustment before you begin, not after. In New York City, the candidate notice must explain how to request one. Elsewhere, the recruiter is the right contact.

Do not use AI to answer for you

Some candidates are tempted to run a chatbot during a live or recorded assessment. Assessment instructions may prohibit this, and answers generated this way tend to be generic. Using AI to practise beforehand is a different matter, and many candidates find it useful for rehearsing timed answers.

Before you reach this stage

An AI interview often follows an automated CV screen. A clearly structured CV with standard headings makes it more likely your experience is read accurately. You can check how your CV parses with the ATS resume score checker, and build a clean version with the resume builder.

After the interview

  • Note what you were asked while you remember it. Automated stages are often reused for later applications to the same employer.
  • If you are rejected quickly, it may be a fixed threshold rather than a judgement on you. Where you have the right to an explanation or a human review, you can request it in writing.
  • If you think your data was misused, raise it with the employer first. In the UK you can then contact the ICO; in New York City, DCWP; elsewhere, the national data protection or labour authority.

Frequently asked questions

Can an AI interview reject me without a human ever seeing my answers?

It depends on the employer and the law where you apply. In the UK, the ICO says people have a right not to be subject to solely automated decisions that significantly affect them and should be able to request human intervention. Ask the employer whether a person reviews responses.

Does AI analyse my facial expressions?

Some older tools claimed to. In the EU, AI that infers emotions has been prohibited in the workplace since 2 February 2025, and the Commission’s guidelines say this includes hiring. Elsewhere, ask the employer whether video is analysed or only the content of your answers.

Do I have to consent to an AI interview?

In Illinois, employers must obtain consent before AI analyses a video interview for an Illinois-based job. In other places, consent rules vary. You can always ask whether an alternative format is available.

Will speaking with an accent lower my score?

Transcription quality can vary, which is one reason bias audits and human review matter. Speak clearly at a steady pace, use a good microphone, and if you are concerned, ask the employer how answers are assessed and whether a person reviews them.

How is an AI interview different from a recorded video interview?

A recorded video interview may be watched entirely by people. An AI interview uses software to ask questions, score answers or summarise them. The same recording can be either, which is why it is worth asking.

Sources

Ready to build your resume?

Start with a sample or a blank page. No sign-up needed. Free to edit; US$1.00 only if you download a PDF.

Build My ATS ResumeExplore Resume Templates