AI IN PSYCHOMETRIC ASSESSMENTS: From Psychological Measurement to Algorithmic Judgment - Hardcover

Shermon, Ganesh

Hardcover

ISBN 10: 199721802X ISBN 13: 9781997218029

Publisher: Ontario Publishing House, Toronto, ON, 2026

Synopsis

AI IN PSYCHOMETRIC ASSESSMENTS: From Psychological Measurement to Algorithmic Judgment How Artificial Intelligence Is Learning to Measure, Interpret and Judge Human Potential What happens when the systems designed to measure human intelligence, personality, aptitude and potential begin to make decisions about how people should be assessed, interpreted and judged? Psychometric assessment has long provided organizations with structured methods for understanding human differences. Intelligence tests, personality inventories, aptitude measures, competency assessments and other psychological instruments have helped recruiters, psychologists, consultants and human resource professionals make more informed decisions about people. Yet the arrival of artificial intelligence is changing the foundations on which these assessments operate. The technology is no longer confined to administering a questionnaire, calculating a score or generating a report. Increasingly, intelligent systems can generate test items, select questions, adapt assessment pathways, evaluate responses, interpret behavioral evidence and produce recommendations that influence consequential decisions about people's lives and careers. AI in Psychometric Assessments: From Psychological Measurement to Algorithmic Judgment examines this emerging transformation and its implications for the future of psychological measurement. It explores the movement from conventional psychometric instruments and computer-based testing towards adaptive algorithms and agentic AI systems capable of planning, acting, evaluating evidence and modifying the assessment process as it unfolds. This is not simply a story of faster testing or automated reporting. It is an examination of how the responsibility for assessing human capability is shifting from human test developers and assessors towards systems that increasingly participate in the design, administration, scoring and interpretation of assessments. The distinction between automation and agency is central to this book. Conventional computer-based systems execute predefined instructions, while computerized adaptive testing adjusts item selection according to a candidate's responses. Agentic AI extends these capabilities by enabling systems to pursue assessment objectives through multiple steps, use available tools, respond to emerging evidence and determine what action to take next within their assigned boundaries. The implications are substantial. An assessment system may increasingly influence not only which questions a candidate answers, but also how the candidate is evaluated, which evidence is considered relevant and how the resulting profile is interpreted. The book establishes the psychometric foundations necessary to understand these developments. It examines constructs, reliability, validity, classical test theory and item response theory, explaining why established principles of measurement remain essential even as the technology changes. It then explores agentic assessment design, automated item generation, computerized adaptive testing, machine learning in scoring and interpretation, and the integration of multiple sources of evidence into candidate profiles. Readers are introduced to the relationship between traditional measurement models and emerging AI capabilities, with particular attention to the question of whether increasingly sophisticated systems can produce assessments that remain scientifically defensible.

Edition details

• Publisher: Ontario Publishing House, Toronto, ON
• Publication date: 2026
• Language: English
• ISBN 10: 199721802X
• ISBN 13: 9781997218029
• Binding: Hardcover
• Number of pages: 310