How America and China Compare on AI
And why the two countries have different definitions of winning.
At least part of Washington's hesitation to regulate artificial intelligence (AI) stems from the desire to ensure the United States maintains its advantage over China in this rapidly evolving field. On the latest episode of FP Live, Matt Sheehan, a senior fellow at the Carnegie Endowment for International Peace, discussed the U.S. versus China AI race.
Sheehan, who lived and worked in China for six years and has facilitated unofficial U.S.-China dialogues on AI, is also the author of "The Transpacific Experiment: How China and California Collaborate and Compete for Our Future."
Sheehan clarified that when it comes to AI, the United States is currently ahead of China in terms of frontier capabilities. This lead is estimated to be six to nine months, meaning that while Chinese models may match U.S. models in capability, it will take time for Chinese models to catch up. However, when it comes to the broader diffusion of AI into the economy, it is unclear which country is leading, as this impact takes time to manifest in data and there is no centralized database to track progress.
The ultimate goal for the U.S. in the AI race is to achieve artificial general intelligence (AGI) or artificial superintelligence (ASI), which would enable a model to perform every task a human can or outperform humans at virtually every task. A crucial step in this process is recursive self-improvement (RSI), where a highly capable model can build its successor, leading to an exponential increase in advancement speed.
U.S. researchers and some companies envision that achieving AGI, RSI, or ASI could grant the United States a decisive strategic advantage, allowing it to maintain geopolitical dominance indefinitely.
Written by urgent.news from Foreign Policy's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
