Some limitations in The Theory of Mind in AI
When discussing the four types of artificial intelligence, the third type comes to mind amid advancements in making machines think like humans, the so-called "Theory of Mind," a machine or system's ability to make decisions equal to that of the human mind. The theory of mind aspect of artificial intelligence constitutes the artificial acumen that a bot, robot or synthetic entity comprised of rule sets and algorithms would be able to perform the same level of thought and emotion as a human. Some of these tasks carry similarities in conducting fully aware and cognitive synthetic attributes designed to communicate and carry human conversation. Theory of mind argues that such conversations would be fluid in language and syntax structure given the programming behind such is written well. It would also have similarities in thought and emotion but lack full self-awareness as this is a complicated function. Unlike humans who have and use memories in conjunction with patterns in life and daily activities, AI would have the ability to recall attributes from the past, but a difficult time making humanistic decisions unless its development were to provide rule sets that provide traceability from the past actions that could alter future decisions. Basic reactive machines power current AI infrastructure in different areas of business, but should self-aware AI improve, the one size fits all method would not work as each use case yields its own characteristics. A good example of reactive memory AI with predetermined rule sets without the inclination of self-awareness would be AI image recognition of object selection. Reactive AI would know what object to select out of any array of objects but not know or feel the reason for its selection.