Building a conscious machine isn’t possible just yet, but it may be in the future.
What would it take for robots to have consciousness?
Humans have three kinds of consciousness. There’s the subconscious, which holds most of our intelligence. The subconscious is what gives us the ability to do things, such as recognize a face in a crowd, without actually knowing how. When it comes to actual consciousness however, researchers say that there are two types. One is capable of maintaining several thoughts at once and enables us to plan for the long term. The other one gives us the ability to be introspective, which enables us to reassess our choices or think about our own mistakes.
Examining human consciousness can help guide the development of conscious machines.
Contemporary artificial intelligence, like DeepMind’s AlphaGo, have processes akin to human subconsciousness. However, existing AI doesn’t have the ability to have the two forms of consciousness. According to a new study, examining these three kinds of consciousness can help researchers answer whether or not machines can ever develop consciousness of their own.
"Human consciousness is not just about recognizing patterns and crunching numbers quickly," said neuroscientist Hakwan Lau, one of the researchers. "Figuring out how to bridge the gap between human and artificial intelligence would be the holy grail."
Researchers designated the kinds of consciousness as C0 (subconsciousness), C1 (perceptual), and C2 (introspection). These are what can serve as markers in a road map that can lead to the creation of a truly conscious machine. If a machine can reach C1, it would be at par with the consciousness of babies and animals. An animal, for example, would know where the nearest source of water is when it’s thirsty and would have the ability to go to that source. Decision making of this kind is possible due to an intricate structure of neural circuits. Most of the activities enabled by C1 is said to take place in the brain’s prefrontal cortex.
Are we ready for self-aware, introspective machines?
C2, however, is what can be called “self-awareness”. Self-awareness is what leads us humans to recognize what we know and what we don’t, which enables us to have curiosity. It is also what drives us to recognize our mistakes and correct them. According to the researchers, there have been robots that have achieved some aspects of C2. These robots can reportedly monitor their own progress in learning problem solving.
So what would it take for robots to have actual consciousness? The researchers say that it may boil down to figuring out the differences between human consciousness and unconsciousness in computational terms. This way, scientists can code the difference between consciousness and unconsciousness into computers.
Get weekly science updates in your inbox!