AI Interview Glossary: 32 Terms Every Candidate Should Know (2026)

The legal term used in NYC Local Law 144 for AI or algorithm-based tools used to make or substantially assist employment decisions about candidates or employees. HireVue, Sapia, and similar interview AI systems are AEDTs. The term is also used by regulatory guidance in other US states.

Illinois state law enacted 1 January 2020. The first US state law governing employer use of AI to analyse video interviews. Requires standalone disclosure, consent, disclosure of specific characteristics evaluated, and allows candidates to request video deletion. Enforced by the Illinois Department of Labor.

A video interview where you record answers to pre-recorded questions and submit them without any live interaction. The most common AI interview format in 2026. HireVue, Willo, myInterview, Spark Hire, and many others use this format. Contrasted with SVI (synchronous video interview), where both parties are present in real time.

Under NYC Local Law 144, candidates can request an alternative selection process instead of the AEDT. The employer must provide one. This right must be described in the LL144 notice sent with the interview invite.

The scoring framework used by AI interview platforms (including HireVue) to evaluate candidate responses. The employer defines a competency, then a rubric of 5 levels, each with specific behavioural descriptors (anchors). The AI matches your answer content against the anchors. Level 5 (exemplary) requires a specific situation, explicit decision-making, trade-offs addressed, and a quantified outcome. STAR-structured answers naturally hit higher BARS levels.

An independent assessment of an AI hiring tool to determine whether it produces discriminatory outcomes for protected groups (by gender, race, age, etc.). NYC Local Law 144 requires AEDT users to publish an annual independent bias audit summary. HireVue had an ORCAA audit in 2021. The term is increasingly used across regulatory frameworks.

The specific rubric document for a given competency, showing the 5-level scale and behavioural anchors at each level. Different employers configure different competency rubrics for the same AI platform.

The document required by NYC Local Law 144 to be published on the employer's website before using an AEDT. Must include impact ratios by race/ethnicity and sex, the date of the audit, and who conducted it. Candidates can ask employers for this URL directly.

AI systems that interact through natural language conversation, either text or voice. Paradox/Olivia and Mya are conversational AI recruiting tools used for screening and scheduling (not deep behavioural assessment). Sapia is a conversational AI used for scored text chat interviews.

The NYC government body responsible for enforcing Local Law 144 (AEDT regulations). Candidates who believe their rights under LL144 have been violated can file a complaint with DCWP. Civil penalties for employer non-compliance range from $500-$1,500 per day per violation.

Comprehensive EU regulation of AI systems, effective in phases from 2024-2026. Hiring AI is classified as high-risk under Annex III. High-risk obligations - mandatory risk assessments, bias testing, human oversight, transparency, candidate right to explanation - take effect 2 August 2026.

Technology that analysed candidate facial expressions during video interviews as a signal of personality or suitability. HireVue used this from mid-2010s to 2021, when they removed it following the ORCAA audit and EPIC complaint. In 2026, no major AI interview platform uses facial expression analysis for scoring. Many prep articles on Google still reference it - they are outdated.

EU and UK data protection law applicable to how employers handle candidate data including video recordings and AI-generated scores. Article 22 of GDPR gives individuals rights related to solely automated decision-making. Candidates in EU/UK can request information about how their data is used and request deletion.

The largest US AI interview platform, founded 2004. Used by hundreds of Fortune 500 companies for one-way recorded video interviews. Scores language content and speech structure via BARS rubrics. Removed facial expression scoring in 2021. Tier-bands candidates before human review.

The EU AI Act classification for AI systems used in hiring, credit, education, law enforcement, and other sensitive domains. High-risk AI systems face the most demanding obligations under the Act: risk assessments, bias testing, human oversight, transparency, and logging. AI interview platforms are classified as high-risk under Annex III, with compliance obligations from 2 August 2026.

A statistical measure used in bias audits to compare selection rates across demographic groups. NYC Local Law 144 requires that published bias audits include impact ratios by race/ethnicity, sex, and intersections. A significant difference in selection rates between groups is a signal of potential bias.

See AEDT. Shorthand for New York City Local Law 144 of 2021, the AEDT law effective from enforcement on 5 July 2023.

The branch of AI that processes and analyses human language. Sapia uses NLP to evaluate written text responses. HireVue uses NLP as part of language content scoring. NLP identifies linguistic patterns, sentence structures, and content signals that correlate with the competencies being measured.

See AVI. A video interview where you record answers to pre-recorded questions without any live person on the other end. The most common AI interview format in 2026.

The independent auditing organisation commissioned by HireVue to audit its AI scoring algorithms in 2021. The ORCAA audit found no evidence of illegal bias and resulted in HireVue removing facial expression analysis. Cited as an example of algorithmic accountability audit practice.

Technology that monitors candidates during online assessments to detect cheating or non-compliance. Talview is the most prominent AI interview platform with built-in proctoring. Features can include environment scan (showing the room to camera), ID verification, keystroke monitoring, and focus detection. Becoming more common in high-stakes assessment contexts.

The question or instruction shown to you at the start of an AI interview response window. HireVue prompts are usually pre-recorded video or audio questions. Sapia prompts are displayed as text in the chat window. Knowing how to read a prompt and identify the implied competency it is measuring is an important prep skill.

A legal obligation for employers to make adjustments to standard processes to avoid substantial disadvantage to candidates with disabilities. Under ADA (US) and Equality Act 2010 (UK), this applies to AI interview stages. Accommodations can include questions in writing, extended response time, text-based alternatives to video, and rescheduling.

In AI interviewing, the framework used to score each competency. See BARS. Each competency has a rubric defining what constitutes level 1 through level 5 responses.

Australian-founded text-only AI interview platform. No video, no audio. NLP on written text responses. Untimed. Used heavily in Australian retail and airline sectors (Qantas, Woolworths) and expanding into US and UK markets. Explicitly detects AI-generated content.

The proportion of applicants from a given group who are selected to advance to the next stage. Used in bias audits to compare outcomes across demographic groups. NYC Local Law 144 requires selection rates and impact ratios by demographic group in the published bias audit.

The response framework used for behavioural interview questions. Situation (specific context), Task (your specific responsibility), Action (your specific decisions and steps), Result (quantified outcome). STAR-structured answers match the BARS rubric structure that AI interview platforms use for scoring, making them the optimal format for AI interviews as well as human interviews.

A live, real-time interview where both parties are present simultaneously. Contrasted with AVI (asynchronous). Most AI interviews in 2026 are asynchronous. Live AI avatar interviews are a synchronous format but remain uncommon in mainstream hiring. Live human interviews recorded by Metaview are synchronous for the candidate.

The practice (common on HireVue) of sorting candidates into tier bands (Top/Middle/Bottom) based on AI scores before human review. In many employer configurations, Bottom-tier candidates are filtered before a human views their video. Candidates are not typically told their tier result.

The agreement you accept when completing an AI interview. Includes data retention policies, consent to AI analysis (on some platforms), and anti-plagiarism terms (Sapia explicitly documents AI-content detection in their terms). Worth reading the specific terms from your invite platform before starting.

Canadian end-to-end recruitment platform including recorded video interviews, live video, reference checks, and AI scoring. Highly customisable employer configurations mean the candidate experience varies significantly. Follow your specific invite instructions closely.

Willo is a one-way recorded video interview platform targeting mid-size employers with volume hiring needs. Short format (10-15 minutes total). Used as an alternative to HireVue for employers who do not need full BARS rubric scoring.