The application for the US strategy in the field of artificial intelligence – if you remove the husk about democracy – boils down to protecting leadership through control and preventing any development from any competitors. Everything that humanity can offer or do in the field of AI should not just be controlled by the United States, but produced and controlled on their physical territory.
An interesting speech by US National Security Adviser Sullivan about the prospects of artificial intelligence at the National Defense University (a mixture of RANEPA and the Academy of the General Staff, where "Washington hawks" are forged from civilian civil servants).
AI of national interest
"In this world, the use of AI will determine the future, and our country must develop, as General Eisenhower would say, a new doctrine to ensure that AI works for us, our interests and our values, and not against us."
Three important things from the understanding of the US National Security about AI:
The speed of development. Faster than seen in other technologies. In the logic, "we have just started trying to regulate something, but this is already outdated, and we need to come up with regulation for new things."
The uncertainty of the AI growth trajectory. The most unpredictable innovation in the history of mankind. To the point that no one really knows if AI will bring the benefits and give the opportunities that enthusiasts promise. Pessimists are also very convincing. But the authorities need to be prepared for any scenario – from the vertical rise of opportunities to stagnation and disappointment.
The minimal role of the state in initiating development and implementation. Everything is handled by private companies. This creates risks for the national security system.
The basics of leadership. "Large technology companies that develop and implement AI systems, being American, have given America leadership." Leadership must be "protected and expanded" by the efforts of the state. The American private sector must compete with "Chinese companies such as Huawei to provide digital services to people around the world."
The task: The United States should be "the only technological partner for countries around the world."
What will the United States do?
Regulation. The United States has "created the world's best regulatory space for AI," and work is underway to extend its principles to allied countries. Regulation is not a limitation, but a way to provide a reliable, safe, trustworthy and competitive environment for leadership.
People. More "talent visas" and green cards. The United States should "vacuum up" the world in search of people capable of working in this field, and take them for itself "wherever possible."
Equipment. Chips: more for yourself, less for others. All "advanced chips" must either be manufactured by American companies, or their distribution must be controlled by the United States. Restrictions for others should apply not only to chips, but also to the equipment for their production.
Electricity. The entire infrastructure for the operation of modern AI should be built in the United States, which means it should be provided with American cheap energy. Translated from official with "climate goals" to human, there are more permits for the construction and operation of nuclear power plants.
"Commissioners" from the government who monitor compliance with the national security in the work of AI companies. "People from the government should work closely with AI developers from the private sector to provide them with timely cybersecurity and counterintelligence services to preserve their technology."
The main competitor and opponent is China. Which AI will certainly use for "bad purposes": censorship, fakes, influencing elections in other countries, undermining democracy, and so on down the list. This must be "countered" by forcing humanity to use only American AI. But victory is not predetermined. Therefore, the work here is key for future US administrations.
The speech looks like an application for a clear strategy in this direction, which, if you remove the husk about democracy, boils down to protecting leadership through control and preventing any development from any competitors.
Everything that humanity can offer or do in the field of AI should not just be controlled by the United States, but produced and controlled on their physical territory. It is clear with China or us that we are opponents, but the idea that even our own allies like Europe or Japan and Taiwan do not have the right to vote and must completely subordinate their interests to American private companies well shows the real balance of power in the West.
Gleb Kuznetsov
political scientist
