Overview
- At a recent Royal Institution lecture Geoffrey Hinton argued that what we call qualia can be described as representational processes and that chatbots with perception, memory and error correction could meet that functional test for subjective experience.
- Hinton illustrated his position by saying a system that can report its perceptions, spot errors and explain misperceptions would qualify as having experience under his account.
- Many philosophers and cognitive scientists say this move does not resolve the ‘‘hard problem’’ of why information processing should produce a first-person feel and they reject equating representation with feeling.
- Hinton warned that AI must be shaped like a child through careful training and cited harmful data examples to show how teaching choices can affect system behavior and safety.
- The debate widens existing AI policy questions because accepting machine experience would change ethical duties, influence alignment work and shift public attitudes toward how humans value thinking and feeling.