There are many questions about artificial intelligence. This topic has long been of interest not only to scientists, but also to writers. Currently, cyberpunk has been developed - a direction that describes the development of artificial intelligence. If we are talking about artificial intelligence (AI), then we refer to the field of computer science. Artificial intelligence is controlled by machines, computers, and mainly software. Machines imitate intellectual activity, therefore it is called an artificial, some kind of cognitive function based on the environment, observation and the learning process.
Key Definitions
Artificial intelligence is a term that consists of two words.
Artificial is something not real and is a kind of fake, because it is modeled.
Intelligence is a complex term. It can be defined in different ways: logic, understanding, self-awareness, training, emotional knowledge, planning, creativity. People have intelligence because they can do all of this. We perceive our environment, study in it and take actions based on what we have learned.
In this case, we are talking about natural intelligence.
First mention
The history of the development of artificial intelligence is quite interesting and began about 100 years ago.
In 1920, the Czech writer Karel Čapek published the science-fiction play “Rossum’s Universal Robots”, also known as RUR. It tells about a factory that creates artificial people called robots. These are living creatures that can be called clones. At first they worked for people, but then they revolted, which led to the disappearance of the human race.
Artificial intelligence in literature and films is an extensive topic. Chapek’s example was to demonstrate the importance and impact of AI on research and society.
First developments
The first studies related to the name of Alan Turing. He was the author of the Enigma code, an encryption machine used in Nazi Germany. His research led to the creation of a theory of computation.
A Turing machine is an abstract machine that, despite the simplicity of the model, can build the logic of any algorithm. Discoveries in neurology, information theory, and cybernetics, together with Alan Turing's research, contributed to the development of the idea of the possibility of creating an electronic brain.
A few years after the Second World War ended, Turing presented his well-known Turing test, which was an attempt to determine the intelligence of a machine. The idea of the test was that a computer is called intelligent if machine (A) and person (B) communicate in natural language, and the second person (C) cannot determine which of the communicators (A or B) is a machine.
The history of the development of artificial intelligence continued in 1956, when the first seminar on this topic was held, and with it the field of AI research was born. Researchers at Carnegie Mellon University, the Massachusetts Institute of Technology, and IBM employees pioneered AI research.
If in the 1960s, the mood of researchers was rather optimistic, then in subsequent years the process slowed down somewhat. In the mid-1970s, interest in AI declined markedly.
After the first attempt to develop AI (AI Winter), artificial intelligence “returned” in the form of so-called “expert systems”.
Expert systems are programs that answer questions and solve problems in a specific area. They imitate an expert in a particular industry and solve problems according to existing rules.
After a series of financial setbacks at the turn of the 1990s, interest in AI fell again.
After many ups and downs, Deep Blue is the first chess computer to defeat world champion Garry Kasparov on May 11, 1997.
Over the past two decades, research in this area is expanding. In 2017, the artificial intelligence development market (related to hardware and software) reached $ 8 billion, and research firm IDC (International Data Corporation) predicts that it will reach $ 47 billion by 2020.
What is AI
The development of artificial intelligence is incredibly fast. If in science fiction, AI is often portrayed as robots with human-like characteristics, in fact it can cover everything from search engine search algorithms to autonomous weapons.
Artificial intelligence is currently defined as a narrow AI, since it is designed to perform a narrow task (for example, only face recognition or only searching the Internet or driving a car). However, as a long-term goal, many researchers consider the creation of a common AI (AGI). While narrow AI can outperform humans in performing a specific task, such as playing chess or solving equations, AGI will outperform humans in almost every cognitive task.
The Importance of Safety Studies
In the short term, the goal of maintaining the influence of AI on society stimulates research in many areas, including economic and legal disciplines, various technical problems associated with verification, security and control. The aspect is becoming increasingly important for the AI system to do what a person wants: to control all the mechanisms, from a car and an airplane to a pacemaker or energy supply system. Another short-term objective is to prevent a destructive arms race in the field of autonomous weapons.
In the long run, an important question is what happens if the search for a common AI is successful and the AI system surpasses people in performing cognitive tasks.
As I.J. Goode said back in 1965, the development of artificial intelligence is a cognitive task. Such a system can potentially undergo recursive self-improvement, after which the human mind will not be able to equal it. Coming up with revolutionary new technologies, such a superintelligence can help eradicate war, disease, and poverty, so creating a common AI can be the biggest event.
However, there is concern that this may be the last event if we do not learn how to reconcile the goals of AI with the goals of a person before he becomes a supermind.
The question still remains relevant whether a common AI will ever be created. Some experts believe that its creation is guaranteed. Although both of these possibilities are real, there is also the possibility that an artificial intelligence system can intentionally or unintentionally cause great harm. Today’s research will help to better prepare and prevent such potentially negative consequences, taking advantage of artificial intelligence and preventing negative consequences.
Is AI a danger?
Most researchers agree that a super-smart AI is unlikely to show human emotions, such as love or hatred, and it cannot become benevolent or malicious. When considering how AI can be dangerous, experts believe that two scenarios are most likely:
AI is designed to do something destructive: an autonomous weapon is an artificial intelligence system programmed to kill. When used improperly, these weapons can easily lead to mass casualties. Moreover, the AI arms race could inadvertently lead to a war of artificial intelligence, which will also lead to massive casualties. This risk can be associated with narrow artificial intelligence, but it grows as the level of improvement of AI and increase in autonomy increase.
AI is designed to do something useful, but it develops a destructive method that ensures the achievement of the goal: this will happen in the absence of complete coordination of the goal of AI and humans, which is actually quite difficult. If you ask an obedient intelligent car to be taken to the airport as quickly as possible, this may result in at least a violation of traffic rules, as the car will do not what you wanted, but literally what you requested. If a superintelligent system is entrusted with an ambitious geoengineering project, it can harm our ecosystem as a side effect, and people's attempts to stop it will be seen as a threat to the task.
As these examples show, the concern for advanced AI is not maliciousness, but competency. Super-intelligent AI will successfully achieve its goals, and if these goals are not agreed with ours, it will become a problem.
Many well-known experts in the field of science and technology have recently expressed concern about the media and open letters about the dangers posed by AI.
The idea of successfully creating a common AI for decades has been considered science fiction. Nevertheless, thanks to recent breakthroughs, many milestones in the development of AI have already been achieved, and experts are serious about the possibility of superintellectual activity in our lives. While some still believe that the emergence of AI at the human level will not happen until a century later, most AI researchers at a conference in Puerto Rico in 2015 suggested that this would happen before 2060.
Since AI can become smarter than any person, we have no exact way to predict how it will behave. We cannot use previous technological developments as a basis, because we have never created anything capable of deliberately or involuntarily outwitting us. The best example of what we might face could be our own evolution. People now control the planet, not because we are the strongest, fastest, or largest, but because we are the smartest. But if we are no longer the smartest, are we sure that everything remains under control? Perhaps, in the case of AI technology, the best way to win this race is not to impede its development, but to accelerate it by supporting security research.
Russia and AI
According to experts, the directions of development of artificial intelligence in Russia are primarily focused on mechanical engineering and autonomous systems. The country is calling for increased action by academia, industry, and the military to develop these technologies.
At present, there are universities in Russia where the best specialists of the country in this field — the Moscow Institute of Physics and Technology, Moscow State University and the Higher School of Economics — help to study the development of artificial intelligence.
Russian artificial intelligence technologies are just beginning to develop. At the same time, the country mobilizes resources to dominate in this area. Russia is financing AI projects aimed at image processing, face, voice and data recognition, voice control, as well as the ability to use information from radar and satellites and information support for weapons.
The development of artificial intelligence in Russia is carried out, in particular, by such companies as Yandex, ABBYY, VisionLabs, N-Tech.Lab, Mivar.
Further development of AI
Until the end of this decade, a sharp increase in the development of AI and its increase in business and society are expected. There is speculation that recent successes in this area will ultimately advance the phase of creating a common AI, and this will be the beginning of genuine autonomy.
According to experts, the market value of the AI industry in revenue in 2015 amounted to $ 5 billion, which was significant for such a novice sector. It is estimated that by 2020, exponential improvements and wider adoption should more than double revenue to $ 12.5 billion.
AI software
The prospects for the development of artificial intelligence are associated with the fact that software companies expand the boundaries of automation, search and social networks. Called the brain of the machine, artificial intelligence is likely to include automation in sectors such as autonomous vehicles and unmanned aerial vehicles. AI software should create additional opportunities for business development and social value.
For example, virtual assistants will offer expert help; smart robots or consultants in the fields of finance, insurance, law, media and journalism will provide instant research or conclusions; in healthcare, AI software will conduct medical diagnostics and provide assistance. The development of an artificial intelligence system in trade helps to optimize actions and eliminates the possibility of irrational costs. Other benefits include a significant increase in the effectiveness of R&D projects by reducing time to market, optimizing transport and supply networks, and improving management through more efficient decision-making processes.
Autonomous driving, although it is in the early stages of development, has also made tremendous progress. And this list is expanding, thereby proving the inevitable impact of AI on our daily lives.
Will human artificial intelligence replace
Unfortunately, technological unemployment is a byproduct of progress.
Mechanized weaving machines reduced the number of artisans-weavers, the tractor left many people without work, and robotics reduced many workers in all types of production. Increasing AI integration will ultimately yield greater productivity in the short term, which will result in less employment.
Concern is legitimate, but during this period new developments of artificial intelligence will not be at such a stage of development that its widespread use will lead to mass layoffs. This technology will continue to be used in relatively niche applications and will not yet reach the critical mass level that threatens global employment.
However, many experts argue that global employment will not disappear. By automating tasks based on analysis, subtle judgments, and problem solving, AI can pose a threat to low-skilled, ordinary jobs in industries such as retail and financial services, and indirectly through the broader automation of the automotive and some other industries. Although it is difficult to predict the exact consequences at this stage, it is estimated that 5% of jobs in these industries are stereotyped by nature. 50-75 million jobs globally, or 2% of the global workforce, are expected to be potentially cut with the advent of AI. Despite the fact that this number is large enough, it loses its significance in comparison with the capabilities that AI will create.
Prospects for the development of artificial intelligence indicate that improving AI and the subsequent surge in productivity will lead to many opportunities for improving the skills of employees and increasing the ability to focus on creative aspects.
In the era of AI, it is planned to increase the number of jobs that require a high level of personalization, creativity or skill - tasks that still require a person to solve.
Developments in the field of artificial intelligence will not only reduce costs by automating processes, but also maximize revenues, helping corporations introduce new categories of products and services.
In the medium term, the industry in question will eventually consolidate, given that a large number of startups are focused on AI. But as industry standards emerge, there will be only a few potential winners. In addition, where there are winners, there are also losers.
Winners:
:
, , , .
, , , , . « » , .
Partly thanks to the development of artificial intelligence systems, companies operating in emerging markets will be able to compete with companies in developed countries, effectively aligning the opportunities for a new industrial revolution.
AI development is currently driven by an exponential increase in computing power and an intelligent device system. Favorable supply factors, such as low computing and storage costs, advanced algorithms, and increased availability of AI-based capabilities, also help develop the necessary conditions for progress.
Many companies are beginning to see him as a creator, rather than a threat to jobs. One of the main obstacles to the development of artificial intelligence technologies is the belief that this will lead to mass unemployment. Some jobs will be replaced by automatic technology. Despite this, there is growing recognition that AI is also creating employment opportunities that did not exist before. Employers are increasingly looking for coders, programmers, and technicians to monitor and support complex, artificially intelligent systems.
Given the benefits of AI, the rapid development of technology is not surprising. Usefulness for business was reflected in the huge number of companies that now use it in one form or another. Nevertheless, there is constant uncertainty about the prospects of the singularity - the moment when artificial intelligence will surpass the human mind. Now that AI can create a new AI, creating a code of ethics for its further development is vital.