How to Slow Down AI? Calls for "Nuclear-Style Consensus" vs. "Blocking China"

[Anchor]

As the debate over whether to slow down the pace of AI continues, opinions are sharply divided on how to actually implement it. In the United States, contrasting solutions have emerged: one suggesting that efforts once made to prevent nuclear war should now be applied to AI, and another advocating for cutting off China.

Here is a report from Washington by correspondent Jeon Byeong-nam.

[Reporter]

Senator Bernie Sanders, a symbol of the U.S. progressive wing, argues that the international community must now apply the same efforts it once used to prevent nuclear war to artificial intelligence.

[Bernie Sanders / U.S. Senator : Like nuclear weapons, artificial superintelligence poses a threat to humanity and must be treated accordingly. During nuclear arms control negotiations where human survival was at stake, the U.S. and the Soviet Union sat down together and precedent shows they agreed on reduction treaties.]

He proposed that the leaders of the U.S. and China negotiate a comprehensive ban treaty on artificial superintelligence, similar to the nuclear warhead and missile reduction agreements made by the U.S. and the Soviet Union in the 1980s.

[Bernie Sanders / U.S. Senator : President Trump should negotiate a comprehensive treaty with President Xi Jinping to temporarily suspend the development of advanced AI and ban superintelligence.]

The solution favored by U.S. conservatives, who prioritize maintaining American dominance in AI, is different.

Steve Bannon, former White House Chief Strategist and an architect of Donald Trump's hardline support base "MAGA," argues that risks can be reduced by cutting China off from the AI ecosystem.

[Steve Bannon / Former White House Chief Strategist : We need to immediately decouple every single part of the AI ecosystem from the Chinese Communist Party.]

He contends that all Chinese students and researchers in the U.S. should be sent back and Chinese investments in the U.S. AI industry must be restricted.

Bannon argued that partisanship must be transcended in the face of the AI threat, calling for a bipartisan agreement between the Republican and Democratic parties.

While consensus on the necessity of AI regulation itself appears to be widening within U.S. political circles, differences in stance remain vast within the industry.

Anthropic CEO Dario Amodei, who sparked the debate on slowing down, proposed a three-step solution: strengthening internal self-inspection, establishing industry safety standards, and fostering international cooperation.

[Dario Amodei / CEO of Anthropic : The relevant industry must come together to create safety standards that apply to everyone. And finally, international cooperation is necessary.]

However, there are also arguments, such as that from Nvidia CEO Jensen Huang, that regulation is entirely unnecessary because companies can simply release products after ensuring safety.

(Photo: Park Eun-ha, Video Editing: Cho Mu-hwan)
※ Please note: This article was translated by AI and may contain errors.