This paper presents a new Local Friendliness (LF) no-go theorem that explores the idea that a system having thoughts is sufficient for it to be considered an observer. This theorem utilizes four metaphysical assumptions, three of which are thought-related, to derive LF inequalities applicable to systems exhibiting "Friendliness". The theorem also relies on assumptions about human-level AI and universal quantum computing.
A proposed experiment using a quantum-mechanical AI to test the Wigner’s Friend thought experiment could have profound implications for our understanding of quantum reality and the nature of observation.