Artificial Intelligence so close yet to far - {Part 1}

#artificialIntelligence, #emotions

The image of the cover appears in October 2022 issue?of?National Geographic?magazine, showing Changa, a chimpanzee at Germany’s Leipzig Zoo through a thermal imaging camera. Researchers from Max Planck Institute found also in chimps a characteristic of humans, when stressed, their noses get cooler. They observed colder noses in chimps that listened to recordings and watched videos of chimps fighting or seeing a person they knew who appeared to be wounded, suggesting the chimps felt empathy.?

This new discovery makes us wonder are machines that are equipped with some kind of intelligence capable of expressing human-like emotions. Or will we create such machines. Do we have the appropriate tools to measure their response to situations that we put them in? Even if Charles Darwin, the Naturalist, Geologist and Biologist, widely known for contributing to the understanding of evolutionary biology, studied in a systematic manner the way humans and animals express their emotions (The Expression of the Emotions in Man and Animals, 1872), only recently researchers, assisted by advances in neuroimaging and neurotechnology, have investigated in such detail the presence of emotions and nonconscious and uncontrollable changes that occur in humans and animals, and have discovered and unfold a world that existed but not have been yet visualized. No similar research effort of such extent has been carried out so far in computing systems and systems that adopt the principles of Artificial Intelligence.

The term Artificial Intelligence was coined in 1956 by the mathematician John McCarth professor at Dartmouth University, at the Dartmouth Summer Research Project on Artificial Intelligence. The main question raised during the event was whether every aspect of learning and characteristic of human intelligence could be simulated by a machine. Was it possible for machines to perform like humans? Natural language processing, computer vision, and neural networks were among the topics discussed during the event, whereas Marvin Minsky, Nathaniel Rochester, and Claude Shannon, were among the participants. This starting point was largely responsible for shaping the Artificial Intelligence research and development framework for the years that followed.

Six years earlier from that Dartmouth Summer Research Project, at 1950, Alan Turing, widely considered to be the father of theoretical computer science and artificial intelligence, publishes?Computing Machinery and Intelligence, where he introduce the concept of thinking machines and the Turing Test. Turing's vision was to built a machine that could learn from experience and to do so it should have the ability to alter its own instructions. The Turing Test, which he introduced is a procedure to determine if a computer system can demonstrate the same intelligence as a human. At 2014, an algorithm that claimed to be a 13-year-old boy named Eugene Goostman managed to convince the 33% of human judges at a Royal Society event that he was actually human. Ray Kurzweil, an American computer scientist, author, inventor, and futurist, predicts that computers will pass the Turing test by 2029. He claims that by then computers will be able to exhibit intelligent behaviour indistinguishable from that of a human.

But first what is the definition of intelligence. This is not an easy task. According to Stuart Russell, a British computer scientist known for his contributions to artificial intelligence, an entity is intelligent to the extent that what it does is likely to achieve what it desires, given that it makes use of whatever it has already perceived. Douglas Hoftstadter, physicist, computer scientist and professor of cognitive science, considers intelligence to include the following: responding to situations with flexibility, understanding ambiguous and contradictory messages from context, recognizing and prioritizing situations based on their importance, finding similarities in seemingly dissimilar situations, finding differences in seemingly similar situations.

Concerning Artificial Intelligence, Russell & Norvig classified Artificial Intelligence Systems into 4 broad categories: systems that think like human beings, systems that think rationally, systems that act like human beings, systems that act rationally. In simple terms, Artificial Intelligence is a field that combines computer science and robust datasets to enable problem-solving. Herbet Simon, was also a significant figure and one of the founding fathers of Artificial Intelligence, stated that Artificial Intelligence can have two purposes. To use the power of computers to augment human thinking, and to use a computer's Artificial Intelligence to understand how humans think. Until today all efforts and outcomes could fall in the first purpose.

Sub-fields of Artificial Intelligence are Machine Learning and Deep learning. Actually, Deep Learning is a sub-field of Machine Learning. Deep learning, the technology making all the noise in the media, is a form of supervised learning. Typically, a Deep Learning model is a neural network with three or more layers, which attempts to simulate the behavior of the human brain. It consists of multiple layers of interconnected nodes, each building upon the previous layer to refine and optimize the prediction or categorization. It represents one of the greatest advances in Artificial Intelligence and for many researchers it will lead to human-level Artificial Intelligence systems. Application of Machine Learning and Deep Learning models will follow in future posts.

Another attempt to categorize Artificial Intelligence is the distinction between Weak and Strong Artificial Intelligence. Weak, also called Artificial Narrow Intelligence, is an Artificial Intelligence that is trained to perform specific tasks. This type of Artificial Intelligence is the one that is all around us at the present time. Chatbots, facial expression recognition, translators, personal assistants, self-driving cars, are among the hot application of Artificial Narrow Intelligence.?Strong Artificial Intelligence could be classified further to Artificial General Intelligence and Artificial Super Intelligence. Artificial General Intelligence, is not present at the moment. It is a theoretical form of Artificial Intelligence that would have an intelligence equal to humans. This type of Artificial Intelligence will have a self-aware consciousness, common sense, background knowledge, abstraction and causality, the ability to solve problems, think, understand, learn. The Artificial Super Intelligence, superintelligence, is the vision of humans to create machines that would surpass the intelligence and ability of the human brain.

Based on what has been said till now, the main question raised earlier, whether machines will be able to express emotions will concern the future systems that will be developed and fall into the category of strong artificial intelligence.

Antonio Damasio, a Portuguese-American neuroscientist one of the neuroscientists that has studied human emotions and feelings and established through his work that emotions play a central role in social cognition and decision-making, appears to be certain that Artificial Intelligence will not in any case express human emotions. For having emotions, this would mean to recreate the human brain, the senses, the body and cognition.

According to Charles Darwin, human beings have emotions as a result of evolution. Human emotions are present to assist the organism to survive. That's the case for Damasio too, who suggests biological mechanisms behind feelings, survival tactics selected and combined in the simple cell organisms of early evolution, a process he defined as homeostasis. Damasio also states that “minds,” should not be thought of being composed by only the nervous systems. Nervous systems corporates with systems of our body, including metabolic, endocrine, immune and circulatory systems.

Can we create such complex systems, with advanced sensors, electronics and mechanical capabilities, that will have the possibility of evolution, in correspondence with the biological analogue and in this way have the possibility of the manifestation of emotions?

Although Artificial Narrow Intelligence is present and plays an often invisible role in everyday life, Strong Artificial Intelligence, the Intelligence that will mimic humans in a way it will be hard to be identified, is still a theory. As Stuart Russell stated we don’t really want intelligent machines, in the sense of machines that pursue objectives that they contain. What we want are machines that are beneficial to us.