We have tried to give a short summary of what intelligence is, and then we have compared the key components which separate the human intelligence versus the artificial one. Different examples of contemporary AI agents have helped us illustrate the pace in which the field is being developed. Parallel to this development, many risks have appeared concerning the future of human civilization. We have also tried to present a solution to the addressed issue, based on our research in this field.
Intelligence has been defined in many ways including: the capacity of logic, understanding, self-awareness, learning, emotional knowledge, reasoning, planning, creativity and problem solving. More generally, it can be described as the ability to perceive or infer information, and to retain it as knowledge to be applied towards adaptive behaviors within an environment or context.
AI (breviation for Artificial Intelligence) – a very recent subject in science, goes even deeper towards this particular phenomenon. Besides studying the human intellect, AI also aims to construct computational machinery capable of human-like or even beyond (human-like) rationality and behavior.
The definitions that were coined to define artificial intelligence are divided in four groups:
Thinking Humanly, Acting Humanly, Thinking Rationally and Acting Rationally. See picture below for popular definitions on these groups.
The principle of how human intelligence works is very similar to the artificial intelligence one. In humans information is transmitted by electrical signals along neurons’ axon. The neuron is the basic working unit of the brain. It is a specialized cell designed to transmit information to other nerve cells, muscle or gland cells. In Artificial Intelligence a similar way of processing information is called the ANN (Artificial Neural Network), which is based on artificial neurons, modelled by the type found in humans.
The ability for humans to learn relies on work of large network of areas in the cerebral cortex which supports our ability to learn and consciously remember everyday facts. AI agents on the other hand have a different way of learning their tasks. Below we have listed the types of machine learning:
Supervised machine learning – With this method the machine can learn from activity in past events and then produce new data. This works with given labeled data which then get processed and predict an output in the future. In the final phase the end result compares with the intended output in order to fix errors.
Unsupervised machine learning – In contrast to the previous example, the machine in this case is given unlabeled sets of data. It cannot provide an intended output, but it can create meaningful structures from unspecified information. (ex. Pictures, Video, Audio etc.)
Semi-supervised machine learning – This method is a cross between supervised and unsupervised machine learning. For this method both unlabeled and labeled sets of data are given, with the labeled ones serving as a training for how the machine will process the unlabeled ones for future predicaments.
Reinforcement machine learning algorithms – With this method the machine interacts with its environment by taking action and discovering whether the result is a failure or success. Through trial and error the machine optimizes it’s path of dealing with the specific task. This is reinforced by rewarding the machine when it performs well, also known as a reinforcement signal.
Energy Efficiency Comparison. The human brain requires 25 Watts to function, whereas a typical machine uses 2 Watts for its learning mechanism.
Universal Comparison. Humans usually learn how to manage hundreds of different skills during their lifetime, whereas AI is usually designed to perform a few amount of tasks.
Multitasking Comparison. Human worker works on multiple responsibilities, whereas the time needed to teach a system on each and every responsibility is considerably high.
Decision Making Comparison. Humans have the ability to learn decision making from experienced scenarios. Even the most advanced robots hardly compete in mobility with a six-year old child, after sixty years of research and development of the field.
State Comparison. The human brain is analogue, whereas computers are digital.
We are providing our paper with two examples of Artificial Intelligence development in the last decade, to help describe the position in which the field finds
Order this paper