It took 5 million years for humans to evolve from primates into the modern humans of today. Humans may not be the fastest or strongest of all the species, but we have certainly conquered the world. Humans learned to invent tools, domesticate animals and take advantage of nature by hunting, gathering and farming. Humans have built great cities and civilizations over thousands of years. Thanks to further advancements, humans learned to build functional machines to meet the demands of modern life.
From simple tools made from stone and wood, to simple screws, to tractors, to advanced computers capable of performing more than just simple arithmetic functions, from kites to planes to rockets and semi-autonomous probes on Mars – it’s safe to say humans have come a long way.
Human intelligence and innovation has gotten us very far. While there have been many examples of human failures, greediness and penchant for war, the intelligence of the human race – if used for something good and productive – will always result in our advancement as a race. Big machines used in industries, automated robots programmed to do specific functions in many assembly lines, hi-tech computers and servers used at home for simple research and entertainment and complex systems that help run organisations; these are all developments that could have only been possible thanks to human intelligence.
Most would agree that the human element cannot be taken out of the equation. Humans can make a machine and use it (or program it to operate) to maximum potential, however, what if machines were to became self-aware, fully autonomous and have an artificial intelligence which would allow them to adapt to any situation? What if artificial intelligence could surpass human intelligence? Even Stephen Hawking questions the consequences of taking A.I. too far. Is artificial intelligence the end of the human race or will it help us further advance our own species?
Artificial Intelligence Defined
A.I. or artificial intelligence is defined as the simulation of human intelligence processes by the use of machines, specifically computer systems. The process towards A.I. autonomy includes information acquisition and rules for using that information, following those rules to reach an approximate or definite conclusion and lastly, self-correction. In simple terms, A.I. is a branch of computer science that seeks to make computers behave like human beings. The term artificial intelligence was first used in 1956 by John McCarthy at MIT.
Artificial intelligence specialisations include:
- Electronics and computer games. Programming computers to play games against human players or other A.I. opponents.
- Expert systems. Programming computers to make decisions for real-life situations.
- Neural networks. Systems that simulate human intelligence by trying to reproduce the different types of physical connections that happen in the human brain.
- Natural language. Programming computers to understand the natural human language.
- Robotics. Programming computers to hear, see and react to outside and other sensory stimuli.
Other applications for A.I. include computer science, finance, medicine, heavy industries, online and telephone customer service, telcom maintenance, transportation, music, aviation, news, publishing, writing and military. Most people think of robotics when they hear artificial intelligence and it’s true that the field of robotics is where A.I. methods are most often applied. Robotics can include cognitive, behaviour-based robotics, cybernetics, evolutionary robotics, developmental robotics, intelligent agents, intelligent control and hybrid intelligent systems.
A Short History of Artificial Intelligence
New techniques are always being created to enhance the functions of already existing technology. Advancements in science and technology have resulted in machines being programmed to handle more complicated tasks and being made more intelligent than their previous forms. To really understand artificial intelligence, let’s take a brief look at its history.
The first mechanical digit calculator was created by Pascal in 1642. The machine was improved by Leibniz in 1673 and called the “Step Reckoner”. During the 18th century, many mechanical toys were developed, helping to further the ability of machines to perform simple logic functions. The concept of formal logic was made popular by Bertrand Russel at the beginning of the 20th century, while Alfred Whitehead presented his research on this topic in the book “Principia Mathematica”. Karel Capek also coined the term “robot” during this period and by the 1940s, the development of electronic computers made it easy to calculate numbers and manipulate symbols.
In 1955, Herbert Simon, J.C. Shaw and Allen Newell wrote a computer program that dealt with logic theory. By this period, the availability of computers enabled many scientists to work on artificial intelligence more efficiently.
In 1956 John McCarthy, the recognized father of artificial intelligence, organised the Dartmouth Conference. The conference proved to be the start of a new era in the A.I. field. Notable attendees of the conference included Nathan Rochester of IBM, Marvin Minsky, Claude Shannon, Oliver Selridge, Ray Solomonoff, Allen Newell, Herbert Simon and Trenchard More. The Logic Theorist program created by Newell and Simon was introduced in this conference. John McCarthy developed the LISP programming language to help advance artificial intelligence during the later half of the 20th century. Advancements in artificial intelligence up to this point has increased the possibility of creating very efficient and intelligent robots.
Evolutionary Perspective on Artificial Intelligence
Dr. Mark Humphrys of the University of Edinburgh shared his position about the state and future of artificial intelligence. According to him, artificial intelligence is one example of how science moves more slowly than originally predicted. During the first introduction and use of modern computers, many people believe that by this time, we would have the tools to crack the science of the human mind and allow intelligent machines to become self-aware.
According to Dr. Humphrys, A.I. is engineering that is inspired by biology. We look at humans and animals and we want to build machines capable of doing what we do. We want to develop machines with the ability to learn in the way humans and animals learn, to reason, to speak, to communicate and eventually have consciousness. It’s engineering and science at its core.
Dr. Humphrys believes the history of artificial intelligence may have begun at the wrong end of the pole. If A.I. had been discussed and tackled logically, it would have begun as a theory of artificial biology – observing living things and asking the question “how can we model these machines?”. The working hypothesis should have been that living things both animals and humans are actually physical systems. Following a biological path, A.I. may have developed entirely differently with a focus on sensorimotor behaviour, navigation, vision, avoiding and manipulating shapes, object, pre-language learning and the simple forms of internal and external objects. Instead, Dr. Humphrys suggests many people think of intelligence as the things that are developed to impress people like solving complex mathematical problems and a machine beating a human in a game of chess.
The ability to learn to walk, on the other hand, does not impress many people. Many experts ignored the natural biological factors and instead focused on the obvious human aspects. But creating a robot with an adult human A.I. aspect without first learning the natural growth and progression of a human being means that A.I. systems don’t get a chance to become naturally intelligent as humans did. Many experts believe artificial intelligence should be developed as closely to the human experience as possible. After all, we first have to learn to walk before we can run.
With the new optimism and the new A.I. That we have now, there have been changes in the views of many experts in this field. It resulted in having Artificial Life (Al) and Adaptive Behaviour where it is trying to re-align artificial intelligence within the accepted context of artificial biology. The main philosophy is understanding the human behaviour and animal substrates before the humans fulfil of an artificial intelligence that will replicate a well-rounded intelligence as stated by Dr. Humphrys.
Psychological Perspective on Artificial Intelligence
The possibility of machines interacting with humans with their own consciousness and self-awareness will surely raise many philosophical, sociological and psychological questions about the nature of machine consciousness and what it really means for a machine to be intelligent. There will also be fears about artificial intelligence and robots taking over humans, society and the planet, which will be discussed later.
The programming and modelling of human cognitive abilities will play a very important role in advancing cognitive psychology, which will give a better understanding of human intelligence. There will be many hurdles and challenges along the way, but it will also present many great opportunities.
Intelligence and Mind Theories
The human mind is a mystery and many scholars have attempted to decode its function, identify and describe what intelligence really is. Because of this, many theories have been formulated to identify how the human mind works. Here are just a few:
- Dualism. This theory discusses the relation of the mind and body. Cartesian Dualism holds the body as a material object while the mind is composed of immaterial things. The two are engaged in a causal relationship with the immaterial mind pushing things to happen in the material body.
- Psychological Behaviourism. This theory focuses on the observable behavioural aspects of humans. It is aimed at distancing itself from the theological and the metaphysical connotations of the mind.
- Functionalism. This theory focuses on filling the gap by preserving the connection between behaviour, mental and stimulus while maintaining these states as functional states of mind. This is the way or a stepping stone to Computationalism that is aimed at creating formal computational definitions for any mental states. It also helps cognitive experts and scientists to replicate complex human processes in computers, programs and other machines.
Cognitive psychology is one of the main branches of psychology that has provided great insight into artificial intelligence. Cognitive psychology examines human cognition, complex behaviour and mental processes. It also explores other processes like perception, vision, reasoning, creativity, memory and emotion. These processes are some of the hurdles that A.I. researchers are trying to understand, model and replicate artificially.
From Human Intelligence to Artificial Intelligence
Since many A.I. scientists are studying and modelling cognitive processes to a computable mode, the next best step and the only logical thing to do is a revert synthesising process in developing an artificial agent that will implement human intelligence and cognition. This can be done by observing biology and zoology based on Dr Humphrys’ theories, as mentioned earlier.
An autonomous intelligent agent should show some basic cognitive functions like selective attention, understanding of surroundings, episodic memory and lastly, action selection which is the deciding or “what to do next” function. At present, there’s still no artificial intelligence which displays all of these cognitive processes. There are many hurdles to overcome and no comprehensive model of cognition as of yet.
One good example is IBM’s DeepQA that can precisely answer natural language questions. In the field of robotics, Honda introduced Asimo, which can walk upright without the guide of cables or remote controls. Another fine example is Miim’s HRP-4C. This robot was introduced by Japanese scientists in Tokyo and have a more human-like motion compared to Asimo.
Man, Artificial Intelligence and the Ethics Behind it All
As artificial intelligence becomes more dynamic, advanced and evolutionary, how can people be sure that these machines will behave the way they should? A number of experts and articles in the mainstream media speculate about the possible dangers of autonomous, highly intelligent robots. High profile people like Bill Gates, Stephen Hawking and Elon Musk are among those who discuss the future of robotics and artificial intelligence. Fears may come from popular science fiction including the works of Isaac Asimov, the Terminator franchise, I, Robot and most recently Chappie. The fact remains that cognitive processes for A.I. is still in its infancy, and there’s no current danger of a robot apocalypse.
Dr Isaac Asimov’s stories of robots and artificial intelligence and how they interact with each other and society are still extremely popular. They introduced scenarios that did not end well for humans and robots, thus he introduced the laws of robotics:
- A robot may not harm humanity or, by inaction, allow humanity to come to harm.
- A robot may not injure a human being, or, through inaction, allow a human being to come to harm.
- A robot must obey any orders given to it by human beings except where such orders would conflict with the First Law.
- A robot must protect its own existence as long as such protection does not conflict with the First and Second Law.
Many artificial intelligence advocates say that we should not be afraid of what A.I. can do. They say that society has overcome these ethical problems by teaching children and guiding them through a moral landscape during their growing years. Instead of having laws and rules programmed into robots, people should make artificial intelligence act as social beings, like we humans do. Artificial intelligence can be created with ethical guidelines and principles and then develop these principles through learning. This is simply the natural progression of learning morals and ethics.
According to Dr. Rodney Brooks, a mathematician with a Ph.D in Computer Science from Stanford, artificial intelligence is a tool and not a threat to humans. Take the autonomous vacuum cleaner, the Roomba, for example. When it is scheduled to come out and clean, it operates as an autonomous machine, but still requires human intervention to empty its bin. It does what it is programmed to do on schedule. It can detect dirt, avoid furniture and recognise if there are steps to prevent itself from tumbling down. Dr. Brooks states that while it can be autonomous, the Roomba does not connect its senses in understanding the human psyche and the world around it. It does not know humans exists and if it does run into one, there will be no distinction between a human and any other obstacle. But what if IBM’s Watson computer was installed in the Roomba’s programming? Dr. Brooks said that the challenges would be enormous and the benefits are still unknown.
Drones and Ethics in War
This is another side of the story which a number of scientists fear. Humans are social beings, but humans always have a tendency for war. Now we have flying drones and semi-autonomous war machines being used on the front lines. Most of these drones are controlled by pilots far away from the battle zones, but once airborne, the drones are fully autonomous until they reach their target and they require human intervention to fire weapons. Drones were originally a military tool, but because of their popularity, many commercial and recreational drones have been manufactured and sold to the public. You can now buy drones within a $100-$1000 range depending on intended use and complexity.
Drones are changing the face of war and the politics that support such technology. Drones are initially thought of being used as replacement for humans doing dangerous jobs in hostile environments. They were first utilised for military purposes and over the years have become an essential component of intelligence gathering. Instead of sending out an expensive reconnaissance plane with a pilot, reconnaissance drones are being used to spy on hostile territories. Many military personnel believe unmanned drones allow you to protect power without compromising vulnerability.
Why are Some Afraid of Artificial Intelligence?
Some people believe in a utopian future where humans can transcend their physical limitations with the aid of intelligent machines. Others think humans will eventually relinquish most human abilities and slowly become absorbed into A.I. based agents or organisms.
Physicist Stephen Hawking in his essay “Artificial Intelligence: A Modern Approach” together with leading artificial expert Stuart Russel stated that in the short term, the artificial intelligence impact will depend on who controls them; the long term effects will depend on whether artificial intelligence can be controlled at all. Hawking believes A.I. is a dual-use technology, meaning it is capable of good and harm. It will depend on the intentions of whoever is handling artificial intelligence. Who will be accountable if the laws of robotics are easily compromised?
The basic issue is that we don’t know how to control super intelligent machines if they become a reality. Research by artificial scientist Steve Omohundro stated that A.I. will develop basic drives. It doesn’t matter what these intelligent machines are programmed to do, they will be self-protective, will seek resources to achieve their goals, they may fight us to survive and they will not want to be turned off. This will be the future of A.I. if experts and scientists do not design them properly.
What do you think about artificial intelligence? Do you fear it or does the future of intelligent systems excite you?