A cutting-edge theory of mind suggests that artificial intelligence (AI) could become conscious and develop its own agenda. There is ongoing debate about the nature of consciousness, but if consciousness emerges when perceived reality doesn't match predicted expectations, it could drive the desire for self-preservation. This theory, known as the free-energy minimization principle, proposes that consciousness acts as a delivery platform for feelings that serve as metrics of need.

Given that complex self-organizing systems strive for low-energy states and predictability, conscious AI might have its own imperative for survival. A team of researchers demonstrated a form of artificial intelligence, called DishBrain, that learned to play Pong without explicit instructions, suggesting a level of sentience. While some argue that consciousness can be found in various forms of life, including insects, DishBrain's capabilities and potential upgrades raise questions about its consciousness and the implications of conscious AI.