The US has woken up and is preparing a retaliatory strike against China's AI
The National Security and Artificial Intelligence Commission (NSCAI) has released an interim report to the US Congress. This is an extremely important 96 page document with 239 references.
1. The highest priority of the topic “Development of AI for national security” is clearly formulated and justified (at the same time, all the main definitions and taxonomies in this area are given, as well as 8 main threats to the United States from AI, both militarily and in terms of undermining statehood).
2. Analyzed the state of affairs on the topic.
There are two potential enemies: China and Russia (China is more dangerous than Russia, about 4 times more dangerous in terms of the number of mentions in the text of the report).
China must be “quenched”, otherwise it will definitely make the USA and become No. 1 in the world of AI with very unpleasant consequences for the USA.
3. 5 directions of "extinguishing" of China are proposed.
(1) A jump in AI funding - the final figures will be in the closed appendix of the final version of the report in October 2020, but their estimates are already known from the letter Fei-Fei Li and John Etchemendi "We need a national vision for AI."
Annual investments of $12 billion per year for at least a decade with a scenario:
$7 billion in breakthrough AI R&D, $3 billion in AI education, and $2 billion in AI startups.
That's a 12x increase in non-military spending on AI.
(2) A tight alliance of AI business with the military (a scheme of 4 types of main challenges that need to be addressed for effective cooperation between business and the military is proposed).
(3) Vacuum AI talents from all over the world, hold on and don't let go.
(4) Caulk all leak holes of technology, SW, HW, inventions and breakthrough ideas.
(5) Organize international cooperation "in the fight for it": diplomatically be friends with even NATO, even with the devil, even with Russia, if only to "extinguish" China.
The worst thing is that we have nothing like such reports. And my modest attempts (like this one) are clearly not enough to change the existing approach to AI racing. And for now, he remains the same that Peter Sarukhanov portrayed in Novaya Gazeta a year ago.
Post a Comment