Is AI the same as making robots

Artificial Intelligence and Consciousness

Can consciousness be recreated?

If one assumes that the brain is a material object that is subject to the laws of physics, then one should also be able to reproduce brain processes in the computer. Provided that the processes have been analyzed precisely enough.

Are consciousness and intelligence nothing more than a process of information processing? The basic processes of information processing between nerve cells are already understood relatively well today. Ultimately, they are biochemically similar to the yes / no decision-making process that also takes place in microchips.

And it has even been possible in the first, simple attempts to connect nerve cords with one another using microchips. If, as some scientists believe, consciousness is nothing more than the sum of our brain functions, then it should also be possible to generate consciousness in robots.

Artificial neural networks that mimic the function of the human brain are supposed to enable robots to learn. In principle, it is about the ability to constantly re-link software (instructions) and knowledge from many nested databases.

A problem can be viewed from different angles and new ways of solving the problem can be created. It is about constantly reevaluating a situation - in other words: about learning. Just like children between the ages of one and two do. In this phase, they too do not yet have a pronounced consciousness.

Parents give them an idea of ​​values ​​and only later, after many experiences and reactions to their own behavior, do children begin to develop values ​​themselves. They reflect on their behavior and thereby begin to develop a pronounced awareness.

Consciousness as a meta-level?

But in principle it is not yet clear today what actually constitutes the human spirit or human consciousness. It is presumably a meta-level that is able to consider and evaluate the information processing processes in the individual brain centers in a superior manner.

Perhaps the cerebral cortex has this function. It receives information from the sensory and motor areas, which are mainly located in the depths of the brain.

On the other hand, many researchers also point out that, from a purely structural point of view, there is no superordinate center in the human brain, but that all areas are interconnected in parallel.

An "I" - the meta-level - could then perhaps arise through the constantly changing, new evaluation of situations that are simultaneously compared with stored knowledge.

And in principle, this would be exactly the process that should take place in robot brains. Some AI researchers believe that there is no need to know exactly what consciousness is and how it is created. One could perhaps learn a lot about this from the evolution of a machine mind.

The computer system "PSI"

Can a machine develop feelings? The Bamberg psychologist Dietrich Dörner tries to use the computer system "PSI", which he helped to develop, to understand the processes in a human brain and to understand how spirit, will and feeling interact.

Together with PSI, he wrote a program in which the hypotheses about the psychological processes occurring in humans are reflected.

PSI is an art being that lives in an artificial world. On an island unknown to it, it has a job to do - it is supposed to collect nucleotides. So it has a motive, and emotions such as anger, anger, pain or joy were programmed into it. PSI can also feel thirsty or hungry and has an instinct for self-preservation.

A neural network enables PSI to learn on its explorations. When PSI does something in its computer landscape, it also keeps a log of its external behavior and the effects on its internal state.

So it reflects and also modifies its own actions. In a second attempt, the neural network is switched off and people have to take control of PSI.

By comparing the behaviors, the psychologists are now trying to find out how the artificial brain and the human brain resolve and process conflicts. The behavior in the different situations is compared with one another.

The researchers hope is to get an idea of ​​how intelligent behavior and consciousness develop. But PSI is just a computer program. It doesn't have a real body and neither does it have real interaction with people.

COG - learnable juggler

Without a body, says the father of the humanoid robot "COG", Rodney Brooks, intelligence cannot develop. Intelligence is only necessary if a being has to assert itself in its constantly changing environment. This is the driving force for the development of intelligence.

And so it is only logical when Rodney Brooks tries to let COG develop in a human environment. COG has eyes, ears and hands. It reacts to language and is able to learn thanks to neural networks. He taught himself to juggle with a metal spiral.

Rodney Brooks does not know whether COG could go through something like an evolution to a consciousness by observing its surroundings and own experiments with objects.

He does not use the word awareness openly in interviews. But it would be a big step if COG could further develop its mechanical skills. He can already drum in time to rock music, and journalists shake hands as well. But he still doesn't have a memory.

Kismet shows emotions

Kismet should also learn to behave like a human. The aim behind this is to develop a robot one day that, as a service robot, can take over care services, for example - while also talking to people and working together.

Kismet has a face that is able to express emotions. He speaks and understands people, shows his first emotional reactions in conversations with people.

Kismet learns from people, just as a child gains new information and begins to imitate it. Based on a basic programming of certain behaviors, Kismet should first react correctly to signals from his teachers and perhaps one day establish links between possible behaviors.

As with COG, however, it is unclear whether and when a form of consciousness, however defined, could develop from this. Wamoeba has a certain awareness of his body.

The humanoid robot at Waseda University in Tokyo has internal sensors. Wamoeba senses its battery voltage and engine temperature. If it runs hot, it switches itself to the back burner and goes to sleep - Wamoeba is programmed to self-preserve.

Visions for the next century

A great father of AI research, Marvin Minsky, said early on: "Spirit is nothing more than a product of mindless, but intelligently nested main and sub-programs."

For the unconventional robotics thinker Hans Moravec it is absolutely clear: "The transformation of the inadequate, biological into a better, digital form of life is imminent - we will all become robots."

By the middle of this century, many implants will make the ephemeral human body more efficient, believes Moravec. And that's just the beginning.

Another - also highly controversial pioneer of the scene - adds one more thing: Ray Kurzweil describes the future of artificial intelligence in a timetable. After that, by the end of this century it will replace man as the crown of creation.

If machines don't even take control of our everyday lives, maybe humans could scan their minds and live on in a robotic body or in the computer.

Chip technology will set the pace for such potential developments. And the development of computing power and thus of electronic brains will continue to grow exponentially. This - so Kurzweil - will open up previously unimaginable possibilities by the end of this century.

Futuristic or realistic - whatever opinions one come to in the face of such visions: They are at the same time terrifying and fascinating.