Eric Schmidt argues against the “Manhattan Project for AGI”
In a political paper published on Wednesday, the former CEO of Google Eric Schmidt, CEO of Scele Alexandr Wang, and Dan Hendorx, the United States should not follow a batch similar to the Manhattan project to develop artificial intelligence systems with “Superhuman” intelligence, also known as AGI.
The paper, entitled “SuperINTelligence strategy“It confirms that an aggressive offer by the United States to control the exclusive artificial intelligence systems can lead to evil revenge on China, and perhaps in the form of an electronic attack, which can destabilize international relations.
“[A] Manhattan Project [for AGI] Competitors are assumed that they will agree to a permanent or pesticide instead of moving to prevent it. ”The participating authors write.
Three very influential characters in the industry of artificial intelligence participated in her authorship The American Congress Committee suggested “the lines of the Manhattan Project” An effort to fund Agi Development, similar to the American atomic bombs in the 1940s. US Energy Secretary Chris Wright recently said the United States is in.The beginning of the new Manhattan project“Artificial intelligence while standing in front of the super computer site alongside the co -founder of Openai Greg Brockman.
The SuperIntellign strategy paper challenges the idea, which many American politics and industry leaders have defended in recent months, that a government -backed program that follows AGI is the best way to compete with China.
In the opinion of Schmidt, Wang, and Hendrix, the United States is in something from the face of AGI that does not differ The aforementioned destruction is mutual. In the same way that global powers do not seek monopolies on nuclear weapons-which can lead to a preventive strike of a deduction-Schmidt and its co-authors argue that the United States should be careful in the race towards controlling very strong artificial intelligence systems.
While the likeness of artificial intelligence systems to nuclear weapons may seem extreme, world leaders already consider Amnesty International a better military advantage. Indeed, and The Pentagon says artificial intelligence helps to accelerate the army’s killing series.
Schmidt and others. Entering a concept they call a malfunction involved in artificial intelligence (MAIM), where governments can proactively disable artificial intelligence projects instead of waiting for arms opponents.
Schmidt, Wang and Hendrycks suggest that the United States convert its concentration from “winning the race to Supertelligence” to developing methods Determine other countries Created by Superintyligent Ai. Participating authors argue that the government must “expand [its] Electronic attack arsenal to disable artificial intelligence threat projects “controlled by other countries, as well as limit the arrival of opponents to advanced artificial intelligence chips and open source models.
Participating authors define the division played in the world of artificial intelligence. There is “Domers”, who believe that the catastrophic results of the development of artificial intelligence are a consecutive conclusion and call for countries that slow the progress of artificial intelligence. On the other hand, there is “ostrich”, who believe that countries should accelerate the development of artificial intelligence and I hope everything will succeed.
The paper suggests a third method: Agi’s development approach that gives priority to defensive strategies.
This strategy is particularly noticeable by Schmidt, which was previously vocal for the need for the United States to compete strongly with China in developing advanced artificial intelligence systems. Just a few months ago, Schmidt issued an article Dibsic was characterized by a turning point in an artificial intelligence race in America with China.
The Trump administration appears to be dead to move forward in developing artificial intelligence in America. However, as participating authors note, America’s decisions about AGI are not in a vacuum.
While the world is watching America paying the maximum of the artificial intelligence agency, Schmidt and its participating authors indicate that it may be wise to follow a defensive approach.
