Glossary — hiring assessment & AI-compliance terms

Adverse impact: When a selection procedure passes one demographic group at a substantially lower rate than another. Commonly screened with the four-fifths (80%) rule and central to bias auditing.
AIO / GEO: AI Optimization / Generative Engine Optimization — practices that make content discoverable and citable by AI answer engines (ChatGPT, Perplexity, Google AI). Achieved with clean semantic HTML, structured data, and clear, factual, well-sourced content.
Assessment blueprint: The set of 2–8 skills chosen for a specific role. hire.center recommends the skills that clear the occupation's importance threshold; the recruiter tunes the list. Billing is one credit per blueprint skill per candidate.
Behavioral simulation assessment: A pre-employment method that places a candidate in a realistic, interactive work scenario and evaluates how they behave — rather than what they claim on a résumé or in a self-report questionnaire. hire.center's core category.
Composure Profile: hire.center's fixed five-trait measure of behavior under pressure — Stress Tolerance, Self-Control, Adaptability, Cooperation, Empathy — drawn from O*NET Work Styles. Scored on every assessment at no cost.
Construct validity: The degree to which an assessment actually measures the psychological construct it claims to (e.g., 'negotiation'), rather than something incidental like verbal fluency or confidence.
ESCO: The European Skills, Competences, Qualifications and Occupations classification — the EU's multilingual framework covering ~3,000 occupations in 28 languages. Used by hire.center for European occupation labels and construct identifiers via the official O*NET crosswalk.
EU AI Act: The European Union's regulation on artificial intelligence. It classifies AI used to evaluate candidates as 'high-risk,' triggering obligations around human oversight, transparency, data governance, logging and documentation.
Evidence-first scoring: Scoring in which every rating must cite supporting evidence. In hire.center, each 1–5 score is paired with a verbatim candidate quote, re-verified server-side; unsupported skills return 'insufficient evidence.'
Four-fifths rule: A rule of thumb for adverse impact: if a group's selection rate is below 80% of the highest-selected group's rate, it flags potential discrimination worth investigating.
High-risk AI system: Under the EU AI Act, a category of AI — including recruitment and candidate-evaluation tools — subject to the strictest obligations because it materially affects people's rights and access to opportunity.
Human-in-the-loop (HITL): A design in which an AI system informs but never makes a consequential decision — a human always decides. Central to hire.center and to meeting EU AI Act human-oversight requirements.
Inter-rater reliability: The extent to which two independent judges assign similar scores to the same performance. A key reliability metric hire.center collects during beta.
NYC Local Law 144: A New York City law requiring that 'automated employment decision tools' undergo an annual independent bias audit, publish a results summary, and notify candidates before use.
O*NET: The Occupational Information Network — the U.S. Department of Labor's free, continuously updated database describing skills, knowledge, abilities, work styles and tasks across 1,000+ occupations, each with numeric importance ratings. hire.center's measurement spine.
Situational Judgment Test (SJT): An assessment that presents realistic workplace dilemmas and measures a person's judgment in responding. Meta-analyses show meaningful predictive validity (around r ≈ .26), strongest for interpersonal and teamwork content.
Transferable skills: Cross-occupational skills — like coordination, negotiation, complex problem solving or time management — that apply across many jobs. hire.center assesses 14 of them, selected per role by O*NET importance.

A plain-language glossary of hiring assessment & AI-compliance terms

From definitions to a real scorecard