Edit | Qingyi

Artificial intelligence technology is closing the name of Damacity Sword on the human head from the beginning of birth. In the game "Saiberpunk 2077" (Cyberpunk 2077), the Super Technology Enterprise is the leader in the world, which controls most countries’ governments and lifesons, and penetrates every corner of people’s lives. It is like a huge behemoth in the night of the night, and the bones are marked with nutrients in the whole planet. "Harassed" in the air shuttle, enjoy the welfare of digital hegemony. The general people can only struggle on the ground that is full of radiation, pollution and killing. True is really positive "Zhumen wine threshing, road is frozen". Technology is leaping speed, but the people are suffering. The city of Night is the epitome of high-tech monopoly caused a morbid society.

We can’t help but feel worried about this? It is necessary to make a pivothip, and may even affect the overall future of the human future. If you control in the hands of a few technological giants, can you avoid potential risks to continue the public, benefit mankind?

Patients are, today’s artificially intelligent "leader" is almost situated in the city of Night City – California, USA. These start-up companies that have been ambitious and expanded are now growing into an unable to shake industry giant. Since 2020, the United States has suddenly upgraded the anti-monopoly of the Internet giants. October last year, the House Anti-monopoly Subcommittee announced a heavy report for the 449-page heavy report – "Investigation of Competition in Digital Markets" )

How to crack down on the growing AI technology monopoly? As strong collision impact on physics, the US Government Committee is currently considering an innovative initiative, which will make revolutionary transformation in the field of intelligence.

The top experts of the Professor of Stanford University, the National Ai Research Resource Task Force, referred to as Nairr is studying how to develop "National Research Cloud, referred to as NRC). Make scholars can get the current computing resources currently mastered by a small number of technological giants – access their cloud data centers, and use public data sets for research, the development of non-commercial AI, injecting a strong intensive agent.

In the mid-1960s, the US government’s investment in public science approached 2% of GDP, and returned the same amazing. Today, any scientific giants in the United States have developed in the shoulders of the Federalch and development, R & D. However, since the Cold War, the US government has decreased significantly on basic scientific research. The article of the China Development Research Foundation pointed out that the United States has a peak period of R & D investment in the 1960s, approaching the federal budget of 12%, and after the proportion, it is only about 4%. According to PROJECT SYNDICATE, US scientific research expenditures have fallen to 0.6% of GDP, and the world is ranked 12th. China is doubled to increase the investment in science, and the scientific research project spending accounts for 1.3% of GDP, and it is twice that of the United States from the perspective of economic scale.

The concept of the National Research Cloud Computing Framework (NRC) is actually very simple, and is the data and computing resources required by the US federal government. As the industry and academic development have become more imbalance, scholars want to get these resources increasingly difficult. For the best way to broaden access channels, the government departments provide funds to scholars’ scientists to support the short-term access to existing cloud data, and build long-term public cloud computing systems.


The news is like a flat thunder, causing a fierce dispute, supporting the sound and boycover. NRC even induced the "stress" response of the scientific and technological promoters and the free marketist, questioned "gunfire" mainly intensive two points:

  1. Technology Jocks suspected that this will increase the capacity of large technology companies to "wrap", and the AI ??technology has damaged the rights and interests of vulnerable groups, so they refuse to accept NRC envision.

  2. Free marketist believes that the government’s role is too radical, "reaching" too long. Academic research is fully able to rely on commercial cloud services provided by private and technology companies.

In the face of opposition, the Three scholars of Stanel E. Ho, Jennifer King and Russell C. Wald are scheduled: "You are wrong."

Stanford University "People-oriented" Artificial Intelligence Research Institute (HAI) is the first to raise the Working Group (NAIRR) in 2020. In the past 10 months, here, the best scientists in the United States, how to study how Nairr, eventually present a "heavyweight" white paper "to build a national artificial intelligence research resource: national research cloud blueprint (Building a national ai study" Resource: a BluePrint for the national research cloud.)

"Our survey results are the most powerful response to opponents."

"The United States needs national research clouds"

"We want to say to those who worry about Ai research in the big technology company: It is already this situation. If it is unable to successfully establish a national research cloud, this will only consolidate private enterprises in the dominant position of private enterprises in artificial intelligence. "

"The best server is in the industry." Deep Learning is an important branch of machine learning in artificial intelligence, and is also the main focal point of the US country study cloud. The code is running requires huge strength to support, and the force is "burning". The birth of an excellent algorithm model requires hundreds of thousands, and even millions of dollars. According to Openai’s information, the calculation of artificial intelligent head companies have soared by about 300,000 times in the past seven years. High costs make the academic community difficult to deal with, only Google, Amazon’s giants offer a data center with billions of US dollars in R & D funds.

Under the double difference of funds and equipment, computer talents have embraced the arms of the company. Ten years ago, the artificial intelligence specialty is the same. But now, they choose the possibility of the industry, twice the academic world! This imbalance will have two serious consequences:

First, the private sector research is subject to the decision-making, supervision and veto of the technology company, and eventually causing research shortcomings. Facebook is dismissed by Facebook to instagram on the internal discussion of young women with harmful problems. Second, the scientific research work of the private sector can only point to a broad application area due to the interests of interest.

Everything is just like the Jeff Hammerbacher said: "That sucks! Our smartest people are thinking about how to make people click on advertising!" Hammerbach is Zuckerberg Harvard’s classmates, Facebook’s data team is created by him.

Undoubtedly, the national research cloud is the best way to solve the core contradiction. Because it can broaden the AI ??access resources other than the enterprise environment. R & D, reviewing AI researchers can "transcend" narrow technical fields (including Natural Science, Social Sciences and Humanities) by the development of the country to study clouds. Under the premise of ensuring privacy, the data of the Earth Observations, Labor Markets and Judicial Systems – Only only a few people have permission access – this will guide the human intelligence to solve more complex and urgent social problems.

For those who depend on the government, such as "enemies", the private sector, such as "friends and relatives", said:

"Only relying on the private sector will only hinder the overall development of the AI ??industry, resulting in slowing innovation, and ultimately damage will be the people’s interest. The public sector is critical to promoting basic research and cost effectiveness. Satellite image as an example, 2008, US Geology Service (US Geology Service) charges $ 600 for each satellite image, with the later data open free of charge, this greatly promotes computer vision in researching global warming, habitat changes, poverty and urban expansion. The application produces 30-4 billion US dollars per year. "

In the statement, in the field of calculation, the federal government has a rich empirical trust in the establishment of tip-scale computing facilities, from the Summit system of the Oak Ridge National Lab (the most powerful supercomputer in the world), and then to the National Science Fund The National Science Foundation invests in high-performance computing networks, working with the private sector to play a powerful role in responding to major events such as COVID-19. The cost of relying on business cloud services is 3 to 8 times the system. The national public cloud is impossible, but it brings feedback, it will make the previous efforts and inputs.

The United States is an urgent need to establish an investment in a pinch can be used at any time, monitoring and supervision of AI systems in the public sector "clan" – the best way to achieve this goal is the National Research clouds. Earlier legacy IT systems continued to plague the US authorities, the Government Accountability Office (Accountability Office) has pointed out that, as of 2016, the US Defense Department is still using an 8-inch floppy coordinated enough to destroy nuclear power in the world. "

"NAIRR the federal government," shuffle "AI areas of opportunity, should not become a child’s play between technology and the only skeptics market commentators, zero-sum game there are no winners, losers!"

In fact under the monopoly of technology, science and technology of these "oligarchs" it will not be always the winner.

Monopoly useless

In July this year, US antitrust law enforcement agencies for the first time to sue mobile application store, Google accused of using the Google Play store to charge exorbitant commissions to developers selling apps, deliberately shielding the other competitors on the Android application store. What is surprising is that one of the few world-scale larger than Google’s companies – Amazon, had become a victim. To know that in antitrust litigation, but Amazon "regulars" on the dock.

"National Research clouds will benefit everyone." The original white paper wrote: "NRC will provide superior cheaper high-end computing resources, secure cloud environment of large-scale government data sets, and the necessary expertise, resources by between in close cooperation to benefit academia, government and industry. AI research in order to expand access to critical resources, NRC will provide research support for basic science of artificial intelligence, artificial intelligence, innovation and popularity, as well as continue to enhance US leadership in the field of artificial intelligence . "

Right now, the technology giant control of the AI’s much unstoppable trend. HAI believes that a healthy and wide variety of sources, inclusive of high innovation ecosystem – National Research Cloud computing framework is the only key to reverse the situation. Break "strangle hold" data barriers will have a profound impact on production and research to achieve a balance of AI.

At present "National Research cloud" is still in the conceptual stage, such as "organizational structure, the implementation body, financial arrangements, cooperation mechanisms, hardware layout, data categories, access privileges" and many have not overcome the difficulties, it is also required strong "data normalization" and other public policies and sustained support resources, as well as a large number of ongoing capital investment.

"Building National Artificial Intelligence Research Resources: National Research Cloud Blueprint" (Building a National AI Research Resource:. A Blueprint for the National Research Cloud) Author:

Daniel E. Ho, William Benjamin Scott and Luna M. Scott Professor of Law, Stanford University’s Artificial Intelligence people-oriented Institute (HAI) deputy director.

Jennifer KingStanford University, people-oriented Institute of Artificial Intelligence (HAI) privacy and data policy researcher.

Russell C. Wald, Policy director Jen-based Institute (HAI) artificial intelligence at Stanford University.

Christopher WanDoctor of Law, Stanford University / MBA students pre-reading, is co-author of "artificial intelligence research to establish national resources," the report, provided research and editorial assistance for this article.

Stanford University "people-oriented" Artificial Intelligence Research Institute (Stanford Institute for Human-Centered AI, referred HAI), a Chinese scientist at Stanford University Li Feifei artificial intelligence and philosophy professor John Echeverria Mandy joint as president, is committed to promoting cross-AI cooperation in the field of science and technology to people-oriented, return of human nature, strengthen research on the social impact of artificial intelligence.

National Artificial Intelligence Research Resources Working Group (NAIRR) : National think tank by 12 academics, politicians and industry experts from the tip of the composition, Stanford University professor Li Feifei, Google cloud Andrew Moore, Allen Institute for Artificial Intelligence Oren Etzioni and so on. AI researchers designed to allow researchers aim to get more government data, computing resources and other tools.


Lei Feng Lei Feng net net