Machine Introspection

Making Robots Conscious of Their Mental States proposes logical self-knowledge, including recognizing ignorance and examining goals. This functional sense of introspection does not establish subjective experience or guarantee truthful self-report.^[1]

Task performance and experience

Assessing Consciousness-Related Behaviors Using the Maze Test evaluates navigation accuracy, while Large Language Models and the Patterns of Human Language Use offers a philosophical account of meaningful language. Behavioral scores and interpretive arguments should not be conflated with direct evidence of subjective experience.