The prospect of building artificial systems capable of thinking and acting like humans has been discussed throughout history. In this era of exponential computing and digitisation, it was only a matter of time before an appealing concept like Artificial Intelligence began to gain traction. But first, what is artificial intelligence?
Definition
The computer emulation of human intelligence is known as artificial intelligence (AI). In other words, it is a field aimed at creating systems that can learn and reason like people, learn from experience, figure out how to solve problems in specific scenarios, compare and contrast facts, and perform logical tasks.
The fact that a system has humanoid hardware and physically acts like one is a field of robotics that differs from Artificial Intelligence in that it does not focus on replicating human thought and reasoning.
It should be noted that, if a system could learn and think like a human, it would have notable advantages over it thanks to its speed and calculation capacity like how a seasoned player of Everygame Poker would have advantages over casual players.
How AI Began, and Some Big Steps
Artificial Intelligence is not a new concept at all. To give an example, in the 17th century the philosopher Descartes already theorized about the possibility of intelligent automata. Of course, it was not until the middle of the 20th century that the subject began to take on relevance.
Alan Turing offered one of the first formal challenges in this subject in 1950, stating that “a system is clever enough if it manages to pass itself off as human when asked by a court.” Surprisingly, this test remains a formidable obstacle to this day.
Years later, the renowned computer scientist John McCarthy coined the term “Artificial Intelligence” at the famous Dartmouth Conference in 1956.
However, what seemed like the rise of a branch of technological research ended in speculation and isolated projects for decades, among other reasons. because it turned out to be a field in which no one dared to seriously invest money.
Despite this, significant advancements were made, such as ELIZA, the first chatbot to use natural language processing (NLP) in 1966; the BKG 9.8 program, which defeated the world backgammon champion in 1979; the first autonomous vehicles to travel significant distances in Paris in 1994; and IBM’s Deep Blue AI, which defeated chess champion Gary Kaspàrov in 1997.
With the entry of the new century and the remarkable technological evolution, AI has become one of the unstoppable trends and the milestones in this field are beginning to be very numerous, from expert systems capable of beating a human in any intellectual activity to virtual assistants able to organize ourselves day to day.
AI in everyday life
Artificial Intelligence can be present in almost every part of the modern world, however, it isn’t necessarily in the shape of opulent virtual assistants that can recognise our speech. Let’s look at some examples, starting with the simplest and working our way up to the most complex:
- Devices in the home. Technology is penetrating our homes in every nook and crevice, from smart thermostats to robotic vacuum cleaners. One of the “fundamental” fields of Artificial Intelligence, home automation, has benefited people for many years.
- Spam filters. It is not one of the most striking AIs, but companies like Google or Microsoft apply a multitude of constantly evolving algorithms in order to detect fraudulent and spam-type emails.
- Personalised ads.They are usually highly trained systems that execute a specific cognitive activity and are based on specialist knowledge. Computers that play chess are a good example.
- Expert agents. They are usually highly trained systems that perform a specific intellectual task and are based on the knowledge of specialists. Computers that play chess are a good example.
- Chatbots. Systems that make intriguing use of NLP (Natural Language Processing) and improve with time; allow for cohesive two-way communication with humans, whether oral or written.
- Video games. Perhaps the most obvious, and hence often overlooked, but it has always been one of the most important sources of AI advancement in the never-ending quest to get “the machine” to behave clearly and convincingly. in a game, whether it’s a racing automobile or an enemy soldier
- Autonomous vehicles. There are many companies, and not only automobile companies, that have jumped on the bandwagon (pun intended) of the intelligent automotive industry, developing systems that process huge amounts of data in real-time to determine the correct trajectory of the vehicle, accident prevention, etc.
- Virtual assistants. It’s the closest thing to a movie AI that we can interact with today. It recognizes our voice, adapts to the way we ask for things and is capable of recommending entertainment according to our tastes. One of the strengths of these technologies is that they have a huge number of users who constantly feed them and help reinforce their learning algorithms. Similar to how we use it on our mobile devices, a virtual assistant can take charge of scheduling meetings, proposing articles of interest, recommending contacts, or accurately tracking an employee’s tasks in the workplace.
- Improved customer service. The use of intelligent chatbots to solve doubts or incidents with customers could be one of the most important improvements in the short term for companies, thanks to the savings that it can lead to in support and customer service.
- Productivity increase. Not only can it detect in which processes there are “bottlenecks” within the company, but it can also go a step further and deduce under what specific circumstances they occur.
- Smart analytics. If properly configured for this purpose, AI could be able to analyze data (both structured and unstructured) and draw conclusions much faster than a person dedicated to “Data Mining”.
- Data prediction. Another area where data prediction may be enhanced is the ability to forecast which consumers will not pay their future bills on time.
- Smart sales and marketing. It is possible that we know that our product sells more in certain months of the year, but perhaps we have not had time to analyze if this occurs in all countries or if it is a generalized fact. If we let ourselves be advised by AI, we could find out in which countries, cities, ages, professional profiles, etc. focus on each era without spending too much time analyzing each one.
- Anomaly detection. It could be controlled if business circumstances are occurring outside of a logical trend, regardless of the department: sales, consulting, technical support, etc.
- Sentiment analysis. Within the recognition of natural language or NLP, there is a branch dedicated to the evaluation of the mood of a person based on the written or narrated text. This allows us to have an idea of the level of satisfaction in the interaction with customers, suppliers, and the general public with respect to our brand.
Artificial intelligence technologies
Artificial intelligence is a broad word that covers a wide range of technologies and topics that are all subsets of mathematical and engineering research. Let’s look at the most important ones, starting with identification systems and progressing to machine learning systems.
- Automatic speech recognition. The goal of automatic speech recognition, which is a discipline about acoustics, is to recognize phonemes in a spoken signal. Speech recognition systems analyze the signal picked up by a microphone in order to recognize the words spoken by the user.
- Natural Language Processing or NLP. While speech recognition focuses on a faithful conversion of speech to text, Natural Language Processing or NLP is a discipline that is more linked to the field of linguistics, and its objective is to understand what the user intends when launching a message, a command, question or statement (whether written or spoken). And also what it expects to obtain, as well as analyze the mood and find subjective patterns in them. In short, it is the field that helps communication (mainly sound and written) between man and machine, and vice versa.
- Visual recognition. Visual recognition is the discipline based on the processing of the image or video signal, with the aim of recognizing patterns, shapes, and in the best of cases, faithfully identifying the different elements in an image.
- Text recognition. Because its major goal is to recognize and identify text in image forms, text recognition might be considered a subset of visual recognition. For this task, OCR (Optical Character Recognition) tools are commonly used.
- Big data. Big Data can be considered, without going into technicalities, a large volume of data. Big Data by itself is not a technology, but having a huge amount of data available (preferably structured) is a vital basis for achieving objectives both in Business Intelligence analytics and in the application of certain Machine Learning algorithms.
- Expert systems. Expert systems are those in which all possible human knowledge about a certain branch has been dumped.
- A classic example is that of systems that play chess, which, starting from a whole collection of moves and strategies that have been entered into their memory, are capable of determining the best move (generally based on decision trees) given certain conditions.
- Robotics. Robotics (whether mechanical or software robotics, such as RPA) encompasses a wide range of devices. Whenever a system or robot shows symptoms of intelligence, for example, by being able to make decisions no matter how basic they are, we will be talking about Artificial Intelligence. Keep in mind that artificial intelligence (AI) does not have to be complex; it can exist at any level, including the most basic, and must be distinguished from machine learning, which is the ability to learn from machines.
- Machine Learning. Automatic learning, often known as Machine Learning, is a branch of Artificial Intelligence that aims to teach a computer to learn and relate knowledge in the same way that a human would. To accomplish so, it employs algorithms capable of recognizing patterns in prior data, as well as new trends like Deep Learning and its neural network algorithms, to make future forecasts.
- Deep Learning. Deep Learning is a subdiscipline of Machine Learning. It is a learning system that is inspired by the functioning of the neural networks of the human brain to process information, with a very complex mathematical basis behind it. Although it is based on experience (either previous data, generated by the environment or self-generated), it is not based on strict indications that determine what is correct and what is not, so that the system can determine conclusions on its own.
- Cognitive Intelligence and Cognitive Services. Cognitive Intelligence is a mix of the technologies outlined above with the goal of building Artificial Intelligence systems that can understand humans. It is the union of visual and sound recognition, reading comprehension, NLP, and Machine Learning to create systems capable of understanding information related to human interaction and responding accordingly. Companies like Microsoft make Cognitive Services available to their customers to be able to extend the capabilities of their applications.
Artificial intelligence categories
Artificial Intelligence is difficult to categorise, and the fact is that it is best to practice categorizing it based on the methods utilized by a particular system. However, based on his ideas, other specialists have attempted to form artificial intelligence groups.
Artificial intelligence can be classified into the following categories, according to computer scientists Stuart Russell and Peter Norvig:
- Systems that think like humans. These systems try to emulate human thought in a quite literal way through artificial neural network models.
- Systems that act like humans. These systems focus on acting like humans; they are more linked to classical robotics and are less flexible.
- Systems that think rationally. When observing, reasoning, and acting, these systems attempt to use human logic. They are not programmed to mimic neuronal action in the brain, but rather to act in a human-like manner in a specific context. Expert agents are an example of this.
- Systems that act rationally. They try to rationally emulate human behaviour, drawing their own conclusions to given environmental conditions. The differential point in these systems is to try to apply rationality to their decisions.
A more common categorisation is one that divides two large groups:
- Weak AI. Although the name may appear to be disparaging, it refers to all current Artificial Intelligence. It is artificial intelligence that is committed to tackling a specific problem or a group of problems in the most efficient manner possible but without the ability to expand to general problems without the appropriate programming. This category includes even the most advanced virtual assistants.
Strong AI. It is Artificial Intelligence capable of equaling or surpassing human intelligence in reasoning and deduction capacity. Today it is a utopia that only exists in science fiction because despite the fact that machines already surpass us in a multitude of capacities (including vision and auditory recognition in some areas), they do not have real feelings or native cognitive capacities, as well as own consciousness and adaptability to any scenario.
David Tobin did his degree in psychology at the University of Edinburgh. He is interested in psychology, mental health, and wellness.
