Do conscious AIs need to have the capacity to feel pain?
How can we be sure that the AI feels pain? The “pain reaction” can be sensor-activated tape. When somebody touches a robot in a certain way. It can say, “It hurts”. This same reaction is possible to create. Using a tape recorder and an electric switch. That means the AI can have multiple reactions that mimic human reactions. So if the bot car has certain sensors and somebody kicks it, that system can say. “Please, don’t kick me”.
Do the AI need to feel pain? That’s a good question, because if we want to create artificial superintelligence, we need to create a machine with consciousness. And that is very dangerous and difficult. Researchers and philosophers don’t agree on what consciousness is or what it entails to be you. Consciousness is being you, but how does that thing form? We say that some part of consciousness is the sum of information that our senses send to the brain. But that’s the first thing that we agree on.
The second thing is that consciousness is the thing that is connected to memory. Our experience modeling our way to react to something. But the thing that we see the world as being at the middle of is a mystery. This thing is consciousness. But how does that thing form? We can make machines that can mimic feelings very easily. When somebody presses the robot’s hand with a certain power. That launches a reaction. Where the robot says, “This hurts”. The fact is that a similar reaction is possible to create.
By using a scale and a cassette recorder. When certain pressure impacts the scale. The hand presses the power switch. That conducts electricity to the recorder that plays the message to people. This thing means that. Robots that mimic pain are easy to make. But if we want to make a conscious robot, we must realize one thing. The robot can mimic human reactions very easily. If the robot is left behind, it can say, “Please don’t leave me”. Or, when somebody kicks the robot's shin, it can say, “This hurts”.
In that case. Reaction is like reflexes. When something happens and there is a match in the system memories, that activates a reaction. If some merchandise’s weight is higher. Than a certain level. The robot can report that the merchandise is too heavy. The human-shaped robot requires the scales because it may carry too heavy things. Or presses too strongly. That breaks the robot or merchandise.
The robot can use its own touch sensor that activates that reaction. Or the surveillance cameras can send that information to a robot. Those reactions are preprogrammed reactions. For a certain situation. Reactions are like tapes that certain actions activate. This doesn’t require deep consciousness.
These kinds of reactions don’t mean that the robot really fears that situation. Another thing is that. If somebody tries to shut down the AI. There can be a program that only authorized persons can perform that operation. The system can only transform to use a secondary power line. That means the AI can refuse to shut down its main unit, even if it's not conscious. Consciousness makes AI dangerous. A creature with consciousness protects itself. That is only one way to see those things. The AI doesn’t need consciousness to refuse to shut itself down. And that means. It’s hard to state. If the system is conscious. Or if the system just follows reactions that were programmed into it.