AI Interview FAQ: 22 Candidate Questions Answered (April 2026)

An AI interviewer is software that conducts some or all of a job interview without a human present. In 2026 there are four main types: text chat (Sapia.ai, Mya), one-way recorded video (HireVue, Willo, myInterview, Spark Hire), live AI avatar (emerging, uncommon in mainstream hiring), and assessment hybrids (Harver, Pymetrics+HireVue). Most candidates in 2026 face either text chat or one-way recorded video.

You receive an invite link, complete a practice question, then answer 4-6 behavioural questions by recording yourself within a time limit per question. One re-take per question is usually allowed. After submission, HireVue analyses your language content and speech structure against Behaviorally Anchored Rating Scales (BARS). Candidates are tier-banded (Top/Middle/Bottom) before human review. In most employer configurations, Bottom-tier candidates are filtered before a human sees the video.

No. HireVue removed facial expression analysis from its AI scoring algorithm in 2021, following the ORCAA audit and the EPIC complaint to the FTC. In 2026, HireVue uses face detection only for video quality checks (confirming you are centred on camera and present). Expressions are not scored. Roughly half of the HireVue prep articles on Google still reference facial expression scoring - they are based on pre-2021 information.

Sapia is a text-only chat interview platform. No camera, no microphone. You type your answers to 5-7 behavioural questions in a chat-like interface. Sapia is untimed, so you can take as long as you need per response. NLP analyses your written text: sentence structure, word choice, behavioural indicators. Sapia explicitly excludes sensitive characteristics from scoring and explicitly detects AI-generated content via linguistic fingerprinting.

On HireVue (the most common platform in 2026): language content (keywords, completeness, rubric match), speech structure (sentence shape, specificity, discourse markers), pacing and speaking rate, and rubric alignment to BARS descriptors. It does not score facial expressions (removed 2021), eye contact, attractiveness, clothing, or background. On Sapia: written text only. NLP on sentence structure, word choice, and behavioural content.

On HireVue, Willo, and similar video platforms: aim for 90-120 seconds. Most prompts allow 2-3 minutes. Leave 30-60 seconds unused rather than padding. BARS rubrics reward specificity, not duration. On Sapia text chat: 150-300 words per answer is the practical range. Too short (one sentence) signals disengagement. Too long (500+ words) dilutes the signal. Aim for structured paragraphs with specific examples.

Using AI to prepare (generate practice questions, refine your STAR stories, improve phrasing in advance) is legitimate preparation. Using AI to generate your submitted answers is different. Sapia explicitly detects AI-generated content and documents this in their candidate guide. On HireVue, BARS rubrics reward specificity and authentic language that AI-generated text consistently lacks. The practical advice: prepare with AI, answer authentically.

Yes, explicitly. Sapia documents anti-plagiarism and AI-content detection in their candidate guide (verified April 2026). They use linguistic fingerprinting and behavioural signals (typing pattern analysis). Direct paste-and-submit of ChatGPT output will typically be flagged. Lightly edited AI output may not always be caught but loses the specific, authentic detail that makes responses score well regardless of detection.

Minimum requirements: wired ethernet or tested wifi (minimum 10 Mbps upload), camera at eye level (stack books if needed - laptop at desk height points the lens at your chin), soft front lighting at 45 degrees (a window or ring light), external microphone or wired earbuds, Chrome or Edge for HireVue, neutral background, all other apps closed. Run a 60-second test recording 30 minutes before the real session.

Look at the camera lens (the tiny dark circle at the top of your screen), not at your own face in the preview window or at a second screen with notes. Looking at the preview window reads as looking down and to the side for the human reviewer. Put a small sticky note beside the camera lens to train your gaze. Aim for 50-70% eye contact with the lens; natural glances down while thinking are fine.

In NYC under Local Law 144, you can request an alternative selection process from an employer using an AEDT - this is different from blanket refusal. In Illinois under AIVIA, you can withhold consent to AI analysis of your video (though the interview may still proceed for human review only). In the EU from August 2026, you have the right to human oversight of automated decisions. In practice, most candidates who refuse AI screening entirely are treated as withdrawing their application.

If an employer uses an Automated Employment Decision Tool and you reside in NYC: you must receive at least 10 business days notice before the AEDT is used; the employer must disclose what characteristics it analyses; they must provide instructions for requesting an alternative selection process or reasonable accommodation; and they must publish an annual independent bias audit summary on their website. You can ask HR for the bias audit summary URL.

The Illinois Artificial Intelligence Video Interview Act (AIVIA), in force since 1 January 2020, requires employers to provide a standalone disclosure notice before using AI to analyse your video interview, disclose what specific characteristics are evaluated, and obtain your consent before running AI analysis. If you are applying for an Illinois-based role, AIVIA applies. If you did not receive a standalone notice, the employer may be non-compliant.

The EU AI Act classifies AI used in employment decisions (including candidate screening and evaluation) as high-risk AI. High-risk obligations - mandatory risk assessments, bias testing, human oversight, transparency, candidate rights to explanation - take effect 2 August 2026. If you are a candidate in the EU for a role governed by EU law, you have the right to meaningful information about automated decisions affecting you from that date.

Yes. ADA (US) and Equality Act 2010 (UK) require reasonable adjustments in hiring regardless of AI use. You describe your functional need, not your diagnosis. Specific accommodations to request include: questions in writing in advance, extended response time per question, a text-based interview instead of video, rescheduling to remove same-day pressure, or permission for additional re-takes. Submit requests in writing before the interview.

Research (Benson et al. 2025, Wiley International Journal of Selection and Assessment) shows autistic candidates' response patterns on algorithmically scored video interviews differ measurably from neurotypical patterns in ways that affect scores. ADHD candidates face timed-response execution pressure; dyslexic candidates face comprehension challenges in text-chat formats. You have legal rights to accommodations. See the neurodiversity guide for specific tactics and accommodation templates.

On HireVue, candidates are tier-banded (Top/Middle/Bottom) before human review in most employer configurations. In many setups, Bottom-tier candidates are filtered before a human watches the video. Some employers apply AI recommendations as filtering; others review everyone manually. You typically will not know which mode your employer uses from the invite email. In NYC, the bias audit summary must describe the selection criteria.

Appeal mechanisms vary by employer. In NYC under Local Law 144, you can request an alternative selection process, which effectively routes around the AI score. In the EU from August 2026, you have the right to an explanation of an automated decision. Outside specific legal frameworks, most employers do not have formal appeal processes for AI screening scores. The practical play is to request accommodations or alternative process before the interview rather than appealing afterwards.

One-way recorded video means you record answers to pre-recorded questions and submit them - no one is present during your recording. Most AI interviews in 2026 are this format. Live AI avatar means a video-call-style experience where an AI-generated avatar conducts the interview in real time, adapting follow-up questions. Live AI avatar interviews exist (Tavus-backed tools) but remain uncommon in mainstream hiring in 2026.

Smile naturally at the camera for the human reviewer who watches the video after the AI scores it. Do not force a fixed smile throughout because you think the AI is watching your expressions - it is not, on HireVue since 2021. A genuine smile at the start and end of your answers sets a warm, professional tone. Natural expression variation during your answer is appropriate. A fixed performative smile throughout reads as odd to the human reviewer.

Data retention policies vary by platform and jurisdiction. Under Illinois AIVIA, employers must preserve records for 4 years and candidates can request video deletion. Under GDPR in the EU, you have the right to request deletion of personal data where it is no longer necessary for the original purpose. Under CCPA/CPRA in California, similar data rights apply. HireVue and Sapia publish data retention policies in their privacy documentation. Request deletion in writing citing the applicable law.

Almost certainly yes, in some form - though the specific modalities will evolve. One-way recorded video was the dominant format 2020-2026; live AI avatar interviews are expected to become more mainstream by 2028-2030. Text-chat platforms like Sapia are expanding beyond their Australian retail stronghold. The regulatory pressure (EU AI Act, expanding US state laws) may shift how AI is used in hiring - more transparency, more human oversight - but is unlikely to reduce AI adoption. The skills this guide covers (STAR structuring, knowing your platform, knowing your rights) will remain relevant.