October 07, 2023

The Cloud Can Solve America’s AI Problems

By: Paul Scharre and Tim Fist

One year ago this week, the Biden administration instituted sweeping export controls on semiconductor chips—the sort used in advanced AI—and chip-making equipment going to China. But the current export controls are too blunt an instrument to be effective over the long term. The ban applies to any buyer in China, even for benign commercial uses that do not threaten national security. It also applies not just to U.S. AI chips and chip-making equipment, but also to chips produced outside of the United States using U.S.-origin technology, software, or equipment. These broad restrictions risk damaging U.S. firms’ market dominance, eroding the sustainability of a key lever for AI governance in the future.

These shifts in the semiconductor industry won’t happen overnight, but the United States needs a strategy for governing AI hardware that is sustainable over the long term.

And current export controls are too leaky to be a reliable method for slowing the proliferation of future high-risk AI systems. Actors prohibited from buying AI chips could still train these systems by smuggling chips or legally accessing cloud computing resources. Or they could simply acquire high-risk AI systems directly through theft or open-source models, adapting them for misuse.

Read the full article and more from Foreign Policy.

More from CNAS

View All Reports View All Articles & Multimedia