Design and implementation of interfaces for human subjects to universal intelligence test

As artificial intelligence are rapidly evolving, we have very little understanding of how to measure or even define the concept of their intelligence in relation to the human one. Current approaches to AI evaluation use task-specific testing. This approach fails to determine how general such intelligence is and how it compares to human intelligence. A better understanding of their relationship would help us not only in developing general artificial intelligence but also in a better understanding of human intelligence. A newly emerging field, universal psychometrics, seeks to address this problem by designing universal intelligence tests. This thesis deals with the design and implementation of an interface for a universal intelligence test. It aims to map the current state of development and implementation of universal intelligence tests and to select an appropriate test for the development of the interface. Implements the interface for the selected test and demonstrates its applicability to human subjects. Based on extensive research, the thesis compiles a checklist that potential candidates for a universal test must meet. It also presents one such candidate of its own, the composite test. In addition, it presents a view of sub-intelligence abilities as vectors, which could help in better computing aggregate test results. It then conducts an experiment to demonstrate the usability of the interface for this test on human subjects. The main result of the work is then an implemented interface usable for broader testing of universal test candidates.