The next master of the world - artificial intelligence
Author:Han Yun
Time:2020/10/01
阅读：5647
1. The birth and development of artificial intelligence Alan Turing was a British mathematician, logician, and cryptographer who had a major impact on the development of computer science. He was 19 […]
1. The birth and development of artificial intelligence
Alan Turing was a British mathematician, logician, and cryptographer who had a major impact on the development of computer science. he at1936 The abstract computing "Turing machine" was proposed in 1999, which formalized the concept of algorithm and is still applied in many theoretical and practical researches.It is generally recognized that Turing's scientific achievements laid the foundation for computer science, especially the theory of artificial intelligence. At1950Published the paper "Computing Machines and Intelligence" in 1999, showing that machines can simulate human behavior and perform logical calculations, such as playing chess.
1956At the Dartmouth Conference in 1999, some scientists, mathematicians and philosophers began to develop the concept of artificial intelligence, which truly ushered in the era of artificial intelligence. It is said to be based on neutrality,John McCarthy chose the name "artificial intelligence". As the conference proposal puts it, "This research is based on the conjecture that any aspect of learning or any feature of intelligence can be described precisely in nature so that it can be simulated by machines."
A new phase of AI development begins with20century80Era. Most notable is Japan's fifth-generation computer system (FGCS) R&D project aimed at enabling computers to perform more difficult computing tasks (Japan, 1984). Although previous computers were implemented on a single central processing unit (CPU) to complete calculations by adding logic units, but Japanese scientists hope to endow computers with new functions. Although the computer had great commercial potential, the project was ultimately unsuccessful: up to1992Years, the Japanese government has invested in the project8.5billion dollars, but no breakthrough in the core technology field (Pan, 2016). Eventually the project ran out of money, and AI faded into obscurity for a while. The only advantage of this failure is to let everyone know that the development of artificial intelligence cannot be separated from the development of innovation and advanced software.
arrive20century90At the end of the 1990s, artificial intelligence research was combined with robots and human-computer interaction. Scientists are starting to talk seriously about artificial intelligence "agents" that have emotions and can express them. This has spurred a new field called "affective computing," which aims to analyze human emotional responses and then simulate them with machines. This also helps to improve the dialogue system (chatbot).
Another breakthrough occurred in21century ago102019, marked by the release of a new graphics processing unit from Nvidia. For the first time, computers have become powerful enough to process large data sets and build complex neural network structures for a range of practical applications. This marks the beginning of accelerated applications of deep learning.
The term "machine learning" is used in2006was first proposed to describe automatic understanding of text. A year later, researchers at Stanford University developedImageNet, a large-scale visualization database for visual object recognition software research. These innovations have led to technologies such as computer vision, speech recognition, and machine translation.
since2010Over the years, advances in hardware and software technology have paved the way for the application of artificial intelligence in everyday life. Powerful computer processors and graphics cards, smartphones and tablets make artificial intelligence programs readily available to a variety of users. For example, Microsoft released thexboxbuilt-inKinect, is the first application3DA gaming device with cameras and sensors that capture body movements.
2011Apple released its voice assistantSiri.2014In 2009, Microsoft released the voice assistantCortana.2015Amazon releases smart speaker and voice assistantAlexa. These virtual assistants use a natural language interface to observe, respond and generalize, and make suggestions to users.
2012In 2009, researchers at Google trained a large neural network to recognize images of cats, including16 000processors. training material is10010,000 untaggedyoutubevideo image.
2016year, googleDeepMindThe alpha dog developed by the company is based on4:1Lee Sedol, one of the best professional Go players in the world. Before that match, no one believed that a single computer program could beat such a good Go grandmaster -- but it happened anyway. This victory once again proved the superiority of artificial intelligence over humans.
2018Year,IBMThe company organizes a human and its artificial intelligence systemProject DebaterLive debate between. both sides only15Prepare debates in minutes, covering topics as diverse as telemedicine and the space program. As a result, the AI system not only listened carefully to the opponent's point of view, but responded convincingly with its own point of view. In another trial of training AI to master human language and thinking, AI-basedProject DebaterTo be able to quote the original text and make jokes about the previous point.
What exactly is the mechanism behind the limitless potential of artificial intelligence? When AI reaches human-level performance, we will enter the era of "general artificial intelligence," in which AI will replicate human intuition to solve problems and make scientific discoveries.In the future, machines will be able to memorize all input data and set patterns, and process large amounts of information quickly. By permanently memorizing large amounts of data, machines will gain the ability to recognize patterns and generalize beyond most ordinary people. A "general artificial intelligence" would respond to problems in its environment fully autonomously.
As applications of artificial intelligence continue to multiply, we are witnessing the birth of a new era — one marked by the emergence of entirely new industries and breakthroughs in medicine, bioengineering, robotics, weapons, and space exploration.
At present, although the term artificial intelligence is well known in the world, people still have a lot of misunderstandings about the concept of artificial intelligence and its actual phenomenon. Pop culture is littered with images of crazy robot uprisings, and this is just one of the horror stories that writers have fantasized about. Of course there are also more practical concerns, such as the impact on employment, which, although well founded, have been greatly exaggerated.
The first, second, and third industrial revolutions all created new general-purpose technologies and rapidly changed the way fundamental problems were solved, such as mass production, transportation, and communication. It has also led to improvements in innovation capabilities and overall productivity. The first industrial revolution, centered on the steam engine, led to the development of coal mining, metallurgy and steel, as well as the construction of railways and tunnels;The second industrial revolution, marked by the invention of electricity, drove the development of telecommunications, the rise of the global oil and gas industry, and brought the telegraph, telephone, automobile, and airplane into the lives of the masses;The product of the third industrial revolution, information technology, has reshaped people's way of life and brought computers, the Internet and automated work procedures.
There is no doubt that the advancement of artificial intelligence will revolutionize human civilization and become an inseparable part of human life. Artificial intelligence is reshaping our world, socioeconomic structures, markets and political institutions, and how it affects our individual choices.
2. The impact of artificial intelligence on human beings
In general, the future of artificial intelligence will bring three major trends to human beings.
The first trend is the emergence of superpowers. Some experts not only predict a new cold war, but also compare the potential of artificial intelligence (especially in the military field) with the atomic bomb.
The second trend is that influence will shift from natural resources and financial companies to technology giants, that is, super companies.
For example, Huawei is a new generation5GA leader in communications technology. Most operating systems and engines, as well as a lot of hardware and software, are powered byAlphabet,Microsoft,IBMand Apple Inc. developed or produced. In the field of autonomous driving and aerospace,AlphabetSubsidiary Google, along with Tesla and SpaceX are leaders in the industry. Companies like Facebook and Tencent shape market trends in social media and communication apps. Amazon, eBay and Alibaba dominate online retail and wholesale.The rapid growth of the super companies reveals a new reality on the other side of the race: Competition for resources is a thing of the past, and competition for big data generated by users is the real competition.
A third, more distant trend points to the famous Project of Gilgamesh. If the plan is implemented, a population of creatures based on superhumans will appear on Earth.Obviously, this relies on the endless advancement of artificial intelligence. Against this background, the most critical trend will be the advent of a new "singularity" as predicted by futurists and famous investors such as Ray Kurzweil and Masayoshi Son - the future of technology will developed beyond human comprehension. Some speculate that this will take place in2050Year.
3. Important business applications of artificial intelligence
Today, the practical application of artificial intelligence has long penetrated into various industries, such as the financial industry, medical image processing, and anti-fraud. Artificial intelligence has begun to replace important elements of traditional industries, while also expanding the potential value in fields such as medicine, law and trade.
For example, in the medical field, artificial intelligence has long been able to analyzeMRIandCTimaging data for disease diagnosis, treatment and monitoring.
In business, AI's ability to analyze data and predict outcomes is a big plus for companies using AI. Any calculation and classification tasks that used to take days or even weeks to complete can now be done automatically by algorithms in less than a second. AI will also hold promise for the rapid modernization of extractive industries, especially in the energy sector.
These technologies, ethics and philosophies are revolutionizing the financial industry: Basic financial services are being automated with the help of robo-advisors, while artificial intelligence is being successfully used to detect fraud, assess and manage risk.
The application of artificial intelligence systems is also changing the transportation industry. For example, a successful transition to autonomous driving depends on building an AI infrastructure. Smart cities are a prime example, where vehicles can be automated earlier and more reliably than in current cities.
No surprise, the AI market is growing by leaps and bounds: Recent forecasts indicate that by2030In 2020, artificial intelligence applications will create up to15Trillion dollar wealth.
Of course, on the other hand, artificial intelligence is also changing the political process, and even the election process. We are all familiar with Brexit and2016example of the US election in 2010. The whole world is talking about these two events, because they lead to another, more important issue, which is that artificial intelligence is changing our world at a pace that government agencies cannot keep up with.
Overall, current AI achievements are both positive and complex. As Dr. Kai-Fu Lee, the former president of Google China and an outstanding investor in artificial intelligence technology, said, "Artificial intelligence is the "electricity" of the new era, and it will change the world. However, while artificial intelligence has benefited all aspects of human activity, it is also bringing about the greatest disruption in history to jobs, industries, and human ways of life, and businesses and governments around the world are still unprepared."
4. The influence of artificial intelligence super companies
Jack Ma, founder and executive director of Alibaba Group, said, "The nature of the digital age: Digital technology is not for self-interest, but for altruism; it is not about empowering oneself, but about creating opportunities for others. Digital technology should not only benefit a few people and countries, but also benefit most people in the world. AI's superiority over humans is not a fundamental cause for concern. The question is whether humans are intelligent enough, because it is not intelligence that determines the future, but human intelligence behind it.
If the industrial age turned humans into machines, the digital age has turned machines into humans. I believe that in the future, machines will become more like machines and humans will become more like humans. Machines cannot replace humans, but they can become partners of humans. Technology is not to make human beings unemployed, but to liberate human hands so that human beings can do more valuable things."
42Group CEO Xiao Peng also said, "Today, artificial intelligence has led to the20Positive sentiment similar to that of the dot-com boom years ago. exist20century90In the 1980s, the development of the Internet triggered an unprecedented technological revolution that completely changed the way people lived—whether it was reading news or managing banks. In the following20Over the past year, we have seen policymakers struggle to regulate the ever-changing global internet platform that enables the rapid exchange of ideas, goods and services for billions of people around the world. "
since1969After computers were first connected to each other in 1999, information technology has gone through three main stages of development.The first phase can be described as a period of decentralized innovation, during which many inventions were implemented according to open source principles.
But in the second stage, the commercial value of information technology applications is discovered, which promotes the regulation and commercialization of research.Later, this led to the creation of a monopoly. The first monopoly was Microsoft Corporation, whose operating systemWindowsandIEBrowsers are bundled with each other.
20Years later, with the birth of online search engines and online trading platforms, new monopolies emerged, such as Amazon, eBay andAlphabet. These companies have achieved success in a very short period of time. Today, Google processes data from around the world every second410,000 requests, executed every day35billion searches processed annually1.2trillion requests.
Sales of digital ads opened the floodgates for cash inflows. New monopoly companies use part of the funds to explore new directions, provide new services, and do new businesses: Google continues to developGmail,ChromeBrowser, Android operating system for handheld devices, Google Earth,Google Analytics, for video sharingyoutube,Waze navigation, and finallyDeepMindArtificial intelligence project.
The third stage of development begins with21At the beginning of the century, with the world's most successful social network Facebook(Facebook)The birth is a symbol. Its original mission was "to make the world more open and connected".To this end, the company has opened up its social network to the world, regardless of nationality, language or personal interests. In addition, Facebook users can upload their own designs and applications to the platform. With these features, the company has attracted a large number of users in a short period of time, with the number of users ranging from2008Year1billion grew to2012Year10100,000,0002019at the end of the first quarter of24100 million.
These figures are staggering. But the real issue is not the number of users, but the ability to influence people with the help of artificial intelligence.Contrary to recent findings in behavioral science and psychology, personal information today includes not only someone's name, age, and social affiliations, but also his or her likes—a self-report at the touch of a "like" button. In the current environment, where the lack of clear legal rules restricts the in-depth analysis of open data, technology giants have been able to continuously improve methods to manipulate people's attitudes, such as directing them to specific consumption decisions.
Surprisingly, the megacorporations never hide their capabilities. This is reflected in the following statement: “Google helps people find what they need, and Facebook helps them decide what they need.” The influence of the megacorps has grown by the day, spreading from one field to another. Needless to say, the megacorps will continue to grow. the world's richest8Individuals have6Individuals are technology company founders whose fortunes come from computer software, telecommunications, e-commerce, social networking, and business analytics.
With the arrival of the age of AI, economic and political systems must undergo dramatic changes, whether willingly or not, intentionally or not. In this era, users will get customized services from Google, Amazon and other super companies. Robots will also be used in their call centers. exist3DIn print shops, these companies will be able to3DPrint anything on the printer, including human organs for future transplants. And virtual assistants, provided by a global network of computers, arrange work and leisure for people.
One possible outcome of this trend is the creation of smart detection systems. This will not only store information about people’s preferences and activities, but also successfully predict behaviour, and even shape the desires and intentions of global consumers.
If not from Alibaba, Amazon, Apple, Facebook, Google, Huawei,IBM, Microsoft and Nvidia and other super companies' products and services, the world will become unimaginable. The founders of these tech giants would go on to capture the world's attention as the world's richest men. As the super giants and tech companies use state-of-the-art AI-based solutions and continue to transform digitally, they will be able to deeply analyze individual perceptions and continue to accumulate influence globally.
5. Asilomar AI Principles
In today's world, nothing is more fraught with opportunity and risk than artificial intelligence. As new applications are discovered, the contours of the future become increasingly complex and unpredictable.
The development of artificial intelligence in the healthcare system, as it is understood, heralds such a future: Underlying diseases are predicted and cured before they are even diagnosed. However, with the same analysis, one can also simulate negative scenarios related to the future. Even if drugs can be developed to treat any disease with the help of artificial intelligence, the possibility still exists that this technology can be used to synthesize powerful poisons. The technology to treat genetic diseases with the help of nanobots could lead to covert interventions in the genomes of entire countries.
The trajectory of virtual assistants looks similarly paradoxical. They have followed people everywhere and understood their beliefs, intentions, desires and emotions better than ever before. All of this takes targeting to a whole new level. While the world is governed by consumption, the development of virtual assistants may confine people in a shrinking spiral of temptation for the benefit of high-tech companies.
The application of artificial intelligence in finance has enabled people to obtain loans in mere seconds, but the terms of the loans depend on the assessment of the borrower's reliability. In such a reality, the importance of credit scores rises to a whole new level, which will form quite strict social norms. Those who want more opportunities will learn to be as predictable as possible and obey the behavioral algorithms of credit assessment. Won't this lead to unprecedented increases in social inequality??
The benefits of applying smart technology in modern cities are undeniable. Safer transportation, less crime, a beautiful environment, and high-quality, personalized social services will certainly increase people's happiness. But the greater the opportunity, the greater the cost of failure. A unified infrastructure management system for a smart city must be protected to the maximum extent against cyber-attacks and technical issues. This is because failure would immediately trigger social anxiety and potentially lead to massive casualties and destruction.
Tech giants and megacorporations will continue to amass influence globally. As they use state-of-the-art solutions based on artificial intelligence and continue to transform towards digital, they will be able to deeply analyze the perception of individuals.
As the battle for supremacy in the field of artificial intelligence intensifies, our world needs new international law for the digital age more than ever to lay the foundations of international society. Since developments in AI are unpredictable, these laws must be rooted in fundamental axioms and flexibly adapted to the rapidly changing realities of the digital age. This is why many, including Elon Musk and Stephen Hawking, have warned that the uncontrolled development of artificial intelligence will pose a threat to the future of mankind. Of course, any autonomous AI system, not just a weapon, should be UN-approved and subject to a strict regulatory framework.
In fact, the most appropriate and logical path to formulating AI regulations is to develop a general legal framework that applies to all such systems. Of course, a crucial aspect is that the development and use of AI systems will have to be completely safe.
What progress has been made in this regard? Written at Stanford University, "2030The 2019 AI in Life report makes the following recommendations, paving the way for legal regulation of AI.
(1) Identify a pathway for building technical expertise in AI at all levels of management.
(2) Remove subjective and practical barriers to advance research on the proper function, safety, confidentiality, and societal impact of AI systems.
(3) Increase public and private funding of interdisciplinary research to study the impact of artificial intelligence on society.
2017A conference on the opportunities and challenges of artificial intelligence held in Pacific Grove, California in 2019 can be considered one of the most important initiatives in this regard. Ray Kurzweil, Elon Musk, Stephen Hawking, Demis Hassabis, Jan LeCun and dozens of internationally recognized scientists and experts in artificial intelligence and related fields participated in this event. event. They adopted the "Asilomar Principles of Artificial Intelligence," which laid the groundwork for ethical research in this field.
(1) The goal of AI research should be to build beneficial intelligence.
(2) When investing in AI, fund research that ensures its beneficial nature, addressing thorny problems in fields such as computer science, economics, law, ethics, and social studies.
① How to make the future artificial intelligence system highly robust so that it will not malfunction and be hacked?
②How to increase prosperity through automation while maintaining people's resources and goals?
③ With the development of artificial intelligence, how to amend the legal system and manage the risks related to artificial intelligence more fairly and effectively?
④What values should artificial intelligence abide by? What should be its legal and ethical status?
(3) Active and constructive communication should be formed between AI researchers and policy makers.
(4) A culture of collaboration, trust, and transparency should be fostered among AI researchers and developers.
(5) Teams developing AI systems should actively collaborate with each other to avoid cutting corners in achieving safety standards.
(6) AI systems should be safe, applicable and practicable.
(7) If an AI system causes harm, there should be a way to find out why.
(8) In judicial decisions, whenever a self-developed system is involved, a convincing explanation should be provided and audited by a competent person.
(9) Designers and builders of advanced AI systems are stakeholders in system utilization, abuse, and actions, and they have the responsibility and opportunity to shape these ethical implications.
(10) Highly autonomous AI systems should be designed to ensure that their goals and behaviors are consistent with human values throughout their operations.
(11) AI systems should be designed and operated in accordance with the concepts of human dignity, rights, freedom and cultural diversity.
(12) Now that AI systems can analyze and utilize data, people should have the right to access, manage and control the resulting data.
(13) The application of artificial intelligence to personal data cannot unreasonably shorten people's actual or perceived freedom.
(14) AI technologies should benefit and empower as many people as possible.
(15) The economic prosperity created by AI should be widely shared for the benefit of all humanity.
(16) Humans should choose how and whether to make decisions on behalf of AI to achieve human goals.
(17) The power achieved through the control of advanced AI systems should respect and enhance, not subvert, the social and civic processes on which healthy societies are based.
(18) An arms race with lethal autonomous weapons should be avoided.
(19) Because there is no consensus, hypothetical upper bounds on future AI performance should be strongly avoided.
(20) Advanced artificial intelligence can represent profound changes in the history of life on Earth, and care and resources should be planned and managed accordingly.
(21) Risks posed by AI, especially those that are catastrophic and existential, require effort, planning and mitigation.
(22) AI systems that recursively self-improve and self-replicate can rapidly increase in quality or quantity, Strict security controls must be obeyed.
(23) Superintelligence should only be developed for the broad sharing of moral ideals, for the benefit of all humanity and not for the benefit of one nation or one organization.
The Asilomar AI Principles address important guidelines that will ensure global security while technology advances. First, there needs to be collaboration between AI researchers and politicians. Clearly, such interactions require the support of an international platform, the size and authority of which should ensure the active participation of all parties involved. The organization should participate in the creation of a global regulatory system to address the possible threats posed by the unfair use of artificial intelligence and maintain international security.
For the country, the society, and the individual, how to welcome the era dominated by artificial intelligence?
We need to recognize that humanity is taking its first steps towards practical use of AI applications and systems. Although extreme events are possible, the probability is extremely small. We cannot use this as an excuse to hold back development and stand idly by.
We must really understand and bear in mind that starting today, artificial intelligence will become the master of the world.
As Masayoshi Son, chairman and president of SoftBank Group, said, "Artificial intelligence opens a new chapter in human history, seeking answers to key challenges and risks. We live in a very interesting time, witnessing the beginning of this global change in history, and at the same time can lead the change."
This article is excerpted from the following books:
Han Yun 