First: The Psychometric Test to be Developed and Graded Using AI
The National Center for Examinations and Evaluation is integrating artificial intelligence tools into the development and writing of exam materials, including test translation and adaptation. The writing section of the 'Amirant' English exam is already being tested using AI, with plans to expand this to the 'Yaelant' exam and computerized psychometric testing in the future. The center emphasizes that AI serves only as an auxiliary tool, and all content remains subject to manual review and approval by professionals.
Artificial intelligence is entering the 'holy of holies': the psychometric test is considered the most critical exam for higher education entry, involving significant investment and rigorous professional oversight to maintain uniform standards. Now, the psychometric test and a series of academic entrance exams, such as the 'Amirant' (English proficiency test), are undergoing significant changes with the integration of artificial intelligence for writing and grading exam booklets.
Artificial intelligence is transforming nearly every aspect of life, and the National Center for Examinations and Evaluation (MALV), responsible for psychometric and selection exams for Israeli higher education institutions, is preparing for this new era.
In recent years, the center has explored using AI for test development (writing questions) and for assessing open-ended responses (grading). It has now been decided to integrate AI into both processes, as well as into test translation and adaptation.
This means that some questions will be drafted using AI: artificial intelligence tools will assist MALV in analyzing questions, estimating difficulty levels, and conducting exam simulations. This will streamline development and enable the creation of new question types. However, MALV emphasizes that AI is intended only as an assistive tool, and professional responsibility remains with humans. To preserve the quality of MALV's exams, any output involving AI will undergo review, editing, and approval by professionals; no AI-generated question will be included in an exam without human oversight.
Furthermore, the grading of certain exam questions (currently in English) will be performed using artificial intelligence. The first application of this is the writing assignment in the 'Amirant' exam, which is now being checked by an AI-based system.
At this stage, the assignment is being processed on an experimental basis, but by next year, AI tools will be used to determine scores for the 'Amirant' exam. In the future, with the gradual transition to computerized testing, the use of such systems is expected to expand to the 'Yaelant' (Hebrew proficiency) exams and to the computerized psychometric test.
'We are not staying behind'
To manage this transition, the institution has established a unit dedicated to integrating AI into the development and evaluation of psychometric and other exams. Dr. Yoel Rapp, CEO of the National Center for Examinations and Evaluation (MALV), stated:
'The technological world is changing at a dizzying pace, and artificial intelligence is penetrating every field, including measurement and evaluation. MALV is not staying behind and is in the midst of comprehensive preparation for the AI era. We are establishing a dedicated professional AI infrastructure that will allow us to streamline development processes and improve measurement and evaluation. However, any professional decision and any question integrated into exams will always undergo human oversight and review.'
Dr. Tzur Karlitz, Deputy CEO and head of the Research and Development division at MALV, added:
'Artificial intelligence is changing how people learn, work, and solve problems. We are investing in research and development today to ensure that future exams are more advanced and adapted to changing needs, while remaining fair, reliable, and valid. This process is occurring across the global standardized testing industry.'
