Is there any analysis of AI focused on real world sensory inputs adding to the training models? I'm thinking of the work of Soul Machines and Baby X. Until AI experiences the real world, we are just an abstraction to it.
This is an excellent point. Too many people anthropomorphize AI, mapping their own human experience (and human needs, wants and emotions) onto what is inert code and hardware. In the process, we create blind spots and assumptions that can lead us astray. It would be a useful corrective to describe how an AI "sees" and "experiences" the real world. I'll put some thought into that and will write something about that in the future.
Is there any analysis of AI focused on real world sensory inputs adding to the training models? I'm thinking of the work of Soul Machines and Baby X. Until AI experiences the real world, we are just an abstraction to it.
This is an excellent point. Too many people anthropomorphize AI, mapping their own human experience (and human needs, wants and emotions) onto what is inert code and hardware. In the process, we create blind spots and assumptions that can lead us astray. It would be a useful corrective to describe how an AI "sees" and "experiences" the real world. I'll put some thought into that and will write something about that in the future.