Xia Yiwengeng from Aofei Temple
Produced by Quantum Bit | Public Account QbitAI



△ Trump and Mattis (REUTERS)

An internal document has just been exposed by the New York Times.

In a memo to Trump, U.S. Defense Secretary Jim Mattis implored the U.S. president to develop a national strategy for artificial intelligence, noting that the United States has not kept pace with China's ambitious AI plans.

And this is just one of the US military's countermeasures.

Despite the slow progress of Trump's push, the Pentagon is not sitting still. Another latest news shows that the US Department of Defense is preparing to invest $1.7 billion (about 11.5 billion yuan) in a five-year AI plan.

The new plan and new organization will be officially launched next month.

Of course, things were not all smooth sailing.

Memorandum and 10 billion investment

In May, U.S. Defense Secretary Jim Mattis sent a memo to Trump imploring the president to develop a national strategy for artificial intelligence.

Citing an article by former Secretary of State Henry Kissinger, Mattis called for a presidential commission in a memo in which he said the United States had been left behind by the ambitious plans of China and other countries.

In this previously undisclosed document, The New York Times reads the urgency that U.S. defense officials feel when facing artificial intelligence. These policymakers and advisers believe that AI may change the future of warfare.

However, this concerned memo will not change the situation immediately.

While government bureaucrats say AI is a U.S. research and development priority and part of the president’s national security and defense strategy, U.S. critics blame Trump for not doing enough so far to develop federal policy on artificial intelligence.



△ The Pentagon

The U.S. military can't wait.

The U.S. Department of Defense is independently advancing research related to artificial intelligence. More than a month after Mattis' memorandum was submitted, the Pentagon announced the establishment of the Joint Artificial Intelligence Center (JAIC). Officials did not disclose how many people will be involved, but the new program will be officially launched next month and will open branches across the United States.

According to the latest data disclosed by people familiar with the matter, the US Department of Defense plans to transfer US$75 million of its annual budget to the newly established JAIC, and will invest a total of US$1.7 billion over five years.

That is to say, the US Department of Defense took the lead in formulating a five-year plan for the development of artificial intelligence.

"One of our greatest national advantages is the innovation and talent in the private sector and academic institutions," JAIC Director Brendan McCord said earlier. "JAIC helps develop our partnerships with industry, academia, and allies."

JAIC development encounters obstacles

The core executor of the five-year plan is JAIC.

On June 27 this year, US Deputy Secretary of Defense Patrick Shanahan announced the official establishment of JAIC. The goal of this center is to enable the various AI projects of the Department of Defense to form a synergy, accelerate the use of AI capabilities, and expand the impact of AI tools. The head of the department, McCord, reports to the Chief Information Officer (CIO) of the Department of Defense.

The artificial intelligence projects of the U.S. military's Joint Chiefs of Staff, combatant commands, Office of the Secretary of Defense, and other departments are all coordinated by JAIC.

In terms of specific responsibilities, JAIC's work involves three major aspects.

First, it will guide the implementation of a series of "national mission plans", which are large-scale artificial intelligence projects that are designed to solve critical challenges. At the same time, JAIC will also serve as a platform to promote cooperation on AI-related projects between parties inside and outside the Department of Defense. There are also ethical, humanitarian, and security issues.

JAIC brings together dozens of AI projects from the U.S. Department of Defense, including the famous Project Maven, which aims to identify people and objects in drone videos.

Google once participated in Project Maven, but the search and AI giant was forced to withdraw from the project due to protests involving thousands of employees.

This is a real problem. Not only Google, but AI developers, especially those in Silicon Valley, are very cautious about working with the military and intelligence agencies.

This was completely unexpected by US Department of Defense officials.

Because since the day Silicon Valley was founded, large American technology companies have been taking on work for the military. The Defense Advanced Research Projects Agency (DARPA) of the United States Department of Defense has been investing in and supporting cutting-edge technology research for decades. Many technological advances and many technology companies have originated from projects sponsored by DARPA.

Ironically, however, to this day, these same tech companies are reluctant to work with the U.S. military.

△ Photo by John F. Williams / US Navy

From the researchers' point of view, there is a fundamental difference between developing unmanned vehicles and robots and developing robotic weapons. They are worried that autonomous weapons will pose an unimaginable threat to humans.

Of course, Google is also a minority. After all, there are more than 20 American companies participating in Project Maven. In addition, small but influential Silicon Valley AI labs such as Fast.ai are also cooperating with the US military to accelerate the development of artificial intelligence technology.

Dealing with the “China Threat”

An important factor that prompted both the Department of Defense and Mattis to turn their attention to artificial intelligence is China's rapid progress in this field.

There are two points that Americans are particularly concerned about regarding China's efforts in artificial intelligence. One is the goal of reaching world-leading levels by 2030 as stated in the State Council's "New Generation Artificial Intelligence Development Plan"; the other is the development of many military-civilian integration projects in China's artificial intelligence field.

The U.S. Atomic Scientists Bulletin analyzed that although China does not have an open "military artificial intelligence strategy", the U.S. government has always paid special attention to China's policy trends in this regard.

Under China's military-civilian integration model, the government, academia and business communities work closely together, and civilian and military AI R&D resources are treated equally. Therefore, the innovation potential of Chinese companies is more easily used to enhance military applications.

Since the first plenary meeting of the Central Military-Civilian Integration Development Commission in 2017, China has put forward higher and higher expectations for the development of military-civilian integration. Tsinghua University has a high-end military intelligence laboratory that emphasizes the commonality between military intelligence and general artificial intelligence, and is called "the core tool for serving the strategy of building a strong country in artificial intelligence."



△ A conference hosted by Tsinghua University with the theme of "Exploiting the advantages of military-civilian integration to promote the development of artificial intelligence"

JAIC is an effort by the U.S. military to find a more efficient path for the development of artificial intelligence in response to competitive pressure.

One of the important tasks of JAIC is to win over researchers from the business and academic communities. Although many employees of Silicon Valley technology companies are resistant to the Pentagon and artificial intelligence scientists are resistant to the military application of AI, the US military has close ties with technology development and even more companies.

A report on Xinhuanet mentioned: 85% of the United States' modern military core technologies are also key civilian technologies, and more than 80% of companies producing military products also produce civilian products.

Therefore, China on this side of the Pacific Ocean also regards the United States as an important reference when promoting military-civilian integration.