What is the AI Manhattan Project? An article on the US-China 21st century "technological arms race"
WASHINGTON— Before President-elect Donald Trump retook office, the U.S.-China Economic and Security Review Commission(USCC)under the U.S. Congress proposed in its annual report released last week that the United States needs to launch an AI research and development program similar to the"Manhattan Project" to compete with China in developing and acquiring"general artificial intelligence(AGI)."
In an interview with the Impulsive podcast in June this year, Trump said, "We have to be at the forefront(of AI). We have to stay ahead of China. China is the primary threat."Given Trump's attitude, people in the American technology industry predict that the new government may support such projects.
AGI can learn, understand, and deal with problems in different fields, just as humans do, not just being limited to a specific task.AGI may also reach or even exceed the performance of human intelligence by continuously optimizing and expanding its capabilities. Therefore, it is also regarded as a technology with great potential and risks and is still being explored. The"Manhattan Project" refers to a large-scale cooperative project between the US government and the private sector during World War II, which ultimately led to the successful development of the atomic bomb.
In the report, the USCC recommended giving the executive branch the power to award multi-year contracts and funding to artificial intelligence, cloud computing, and data center companies, integrating the power of government, businesses, and scientific research institutions to accelerate breakthroughs in AGI, and establishing "America's leadership in the field of AGI."
After the USCC report was released, Vinod Khosla, the world's king of"technology" investment and one of the four largest venture capital firms in the United States, wrote in Bloomberg News that artificial intelligence is about to take a huge step forward and has a profound impact on the United States and the world. He called on the Trump administration to work hand in hand with the technology industry.
“The Trump administration and the tech industry need to work together now to prepare for this technological revolution, as it will shape everything from America’s economic power to its competitive advantage over China in computing,” he wrote, adding that his investment company, OpenAI, recently released a system that can reason answers rather than rely solely on predictions. Other leading AI companies around the world are also developing similar systems.
From policy planning to technology research and development, the United States and China, the two largest economies in the world, are competing in different ways to dominate this future technology. Experts point out that this competition is not only technological but also plays a vital role in the two countries' economic competitiveness and national security, and it even affects global values and the future direction of human society.
This article will briefly introduce the AIManhattan Project and the current competition status between the United States and China in this field.
What exactly is AGI?
Artificial intelligence is divided into narrow artificial intelligence(ANI)and general artificial intelligence(AGI).
AGI is an intelligent system that can think flexibly, learn autonomously, reason automatically, and solve problems innovatively like humans. Unlike the current narrow artificial intelligence that focuses on specific tasks, AGI is versatile and can be freely applied in multiple fields without explicit guidance from human designers.
The Importance of AGI
AGI has disruptive potential that could completely change human society's operations and as a core driver of the"next industrial revolution."
For example, in the financial field, AGI can perform financial analysis and understand economic theories across fields, predict the impact of geopolitical risks on the market, and independently formulate investment strategies.AGI can"think" about investment logic like humans rather than rely solely on historical data. For example, in cancer research, AGI may discover treatment paths that human scientists have not yet noticed, and so on.
In the military, AGI is a technology that can"change the game's rules."It can optimize battlefield command through autonomous decision-making, enhance the precision strike capability of intelligent weapons, strengthen network security and attack efficiency, and improve intelligence analysis and logistics management accuracy. However, the military application of AGI is also causing concerns about the risk of loss of control and ethical challenges. Some people worry that this technology can bring the"killer robots" in movies into real life and make unmanned warfare rampant.
Jacob Helberg, a member of the USCC committee, told the Washington Post that AGI would bring about a"complete systemic shift" in military capabilities.
“It’s hard to overstate the profound changes that are coming.OpenAI’s new reasoning models can answer complex questions equivalent to or even slightly better than those of high-level PhD graduates. It will accelerate scientific research, increase economic output, help companies discover new sources of revenue, and increase employee productivity, ”Khosla wrote in the Bloomberg article.“Users worldwide can also obtain professional services from doctors, scientists, accountants, designers, structural engineers, software engineers, etc., driven by artificial intelligence at no or low cost.”
“History tells us that countries that can first exploit periods of rapid technological change are often able to have a significant impact on the global balance of power,” Helberg told Reuters. “China is racing full steam ahead towards AGI. . .We must take this threat very seriously.”
However, AGI does not exist yet and remains a topic of controversy and speculation in the scientific community. Some experts believe that with the rapid advancement of technology, the birth of AGI may be just around the corner. In contrast, others believe true AGI may not be achievable due to insurmountable ethical, technological, and philosophical challenges.
Current AGI technology prototype
Although AGI has not yet been realized in the United States, some existing technologies have already shown some potential for AGI:
China currently has some products and technologies that have the characteristics of AGI prototypes. For example, Baidu's Ernie Bot has the capabilities of natural language processing, logical reasoning, and cross-domain tasks; Alibaba's Tongyi Qianwen can process various forms of input such as text, voice, and images, support multi-domain tasks, and show specific general capabilities; Huawei's Pangu big model also performs well in cross-domain task processing; SenseTime's multimodal AI system combines vision, voice and natural language for perception and decision-making in complex scenarios; and so on.
US-China AI Competition
The USCC report stated that as of June 2024, 50 companies in China were developing AI models, while only a few large companies in the United States were engaged in the research and development of artificial intelligence models.
However, on November 21, the Associated Press reported, citing the latest index released by Stanford University, that the United States leads the world in AI technology development, surpassing China in research and other important AI innovation indicators.
There is currently no universally recognized standard to measure global AI leadership. Still, this Stanford University index attempts to rank the AI industry by calculating its "vitality" in multiple dimensions, including the scale of research and investment and the sense of responsibility in preventing potential harms of the technology.
"The gap between the U.S. and China is widening," Ray Perrault, director of the Stanford AI Index, told the Associated Press. At least in terms of both company creation and company funding, the U.S. is investing more."
The Stanford report shows that the United States leads in private AI investment, reaching $67.2 billion last year, while China only spent $7.8 billion.
The United Nations intellectual property agency said earlier this year that China's patent applications in generative AI far outstripped those of other countries. The Stanford ranking sees this as an indicator of China's strong growth in AI innovation but believes this is not enough to put China in a leading position. The report pointed out that"China's focus on developing cutting-edge AI technology and its increasing R&D investment has made it an important force in the field of AI."
However, according to a report by The Harvard Crimson on November 19, former Google CEO Eric E. Schmidt said at a forum held by Harvard University's Kennedy School of Government that the United States is falling behind China in the race to develop more powerful artificial intelligence.
The newspaper pointed out that Schmidt's judgment showed a change in his position on this issue. Just last October, he said at the same forum that the United States is ahead of China in developing artificial intelligence.
The Harvard Crimson reported that Schmidt pointed out at the forum that in addition to having excellent engineers, powerful chips, and enormous data resources like the United States, China also benefits from more abundant electricity to support the training of AI models, and the Chinese government's subsidies for AI-related companies.
The previous Trump and Biden administrations have introduced measures to restrict China's access to advanced chips and related technologies from the United States, limiting China's access to advanced chips from the United States and its allies. However, Schmidt said Chinese companies have"caught up" with the United States in hardware development by developing their powerful chips.
Khosla, one of the four largest venture capital firms in the United States, also mentioned in his article the starkly different strategies between the United States and China in the AI competition."China's approach is characterized by centralized planning, with direct state funding to support specific AI projects and the establishment of national computing centers, "Khosla wrote."In contrast, the United States relies mainly on private companies to promote the development of AI, many of which adopt open-source models. The U.S. government has relatively limited support, and only recently has it imposed some export controls on the industry."
The AI competition is also a contest of values governance models
In addition to strategically outperforming the United States and its allies by investing heavily in cutting-edge AI technology, China is also seeking to use new AI-driven surveillance and repression techniques to strengthen its illiberal governance model—not only at home but also exporting it worldwide.
“The United States and China are competing not only for leadership in artificial intelligence(AI)innovation but also for whose values guide global AI adoption,” the Atlantic Council, a Washington think tank.“China has continued its strategy in other areas of AI technology. It takes a multi-layered approach to promoting AI development in the global South, ultimately serving its economic and diplomatic interests.”
The think tank's report pointed out that through initiatives like the"Digital Silk Road," China has made large-scale investments in technology and infrastructure projects. Chinese technology companies such as Huawei, ZTE, and SenseTime have provided tens to hundreds of millions of dollars in financing and investment for multiple digital infrastructure projects, including fiber optic cable construction, hardware equipment procurement, surveillance camera installation, and artificial intelligence applications for public sector digitalization. In addition, China is also promoting its AI governance model in international forums such as the United Nations, the Group of 77 (G77), and the BRICS, in part to weaken the influence of Western digital governance models.
“The competition between China and the United States in artificial intelligence and emerging technologies will ultimately have broad impacts beyond the digital realm. The core values of free and open societies are facing challenges, and those countries and alliances that can maintain their advantages over the long term will have greater advantages in terms of economic and national security,” the Brookings Institution, a Washington think tank.“Fortunately, the United States can work with allies and partners in democratic countries to take joint actions to safeguard the values of democracy and freedom in the age of artificial intelligence.”
Trump's new administration's attitude toward the US-China AI competition
In February 2019,then-President Trump signed the United States' first executive order on AI, proposing a strategy to promote the development of AI technology from a national strategic level to ensure that the United States leads the world in this field. In December 2020, the U.S. Congress passed the bipartisan National AI Initiative Act, which includes funding for non-defense AI research and development.
In December 2020, Trump signed another executive order titled"Promoting Trustworthy AI in the Federal Government, "establishing guiding principles for federal agencies to adopt AI technology.
In the Republican Party platform for the 2024 US presidential election, the Party called for repealing Biden’s 2023 Executive order on AI, arguing that the order excessively restricted the development of AI technology.
Trump has called the competition between the United States and China in artificial intelligence" a race that will change the world."He said the large amount of electricity generated through his energy policy would help American technology giants win the arms race.
“During the campaign, Trump and his opponent, Vice President Kamala Harris, made clear that they recognize the importance of ensuring the United States wins the global AI race. Many AI-related bills in the last Congress reflected bipartisanship. AI policy—especially regarding China—will likely be one of the few areas of bipartisan consensus during a Trump presidency,” Khosla predicted.