Can Artificial Intelligence do Autonomous Research?: An Example of Using the Mitchell Criteria

Nieto-Chaupis H. · 2024 IEEE 4th International Conference on Electronic Communications, Internet of Things and Big Data, ICEIB 2024 (2024) · DOI: 10.1109/iceib61477.2024.10602726

Resumen

Nowadays, media and scientific literature have shown capabilities of Artificial Intelligence (AI) to carry out autonomous activities with the same efficiency as a human being are presented. It is clear that AI can support humans in basic scientific research. Thus, one expects that computers can carry out activities ranging between the simplest and the most complex ones. Thus, it is interesting to know if a computer can conduct scientific and technological research, but in an independent manner. As a computer does a typical job, one can wonder about the concrete capabilities of AI to do research independently by using Machine Learning (ML) algorithms and mathematical schemes. Applying Mitchell's criteria in conjunction with well-known mathematical methodologies, a pseudo-code was built and discussed as to explore concrete capabilities of AI to do original research.

Instituciones: Universidad Autónoma del Perú