White House Can't Decide If It Loves or Hates Open-Source AI

The New York Times reports that the Trump administration is caught in a genuine contradiction over how to handle open-source AI — the category of models that anyone can download, modify, and deploy for free. On one hand, these models are central to American competitiveness; homegrown open-source AI gives startups and researchers an edge without requiring billion-dollar compute budgets. On the other, those same freely available models have become a preferred shortcut for Chinese AI labs seeking to close the gap with U.S. frontier systems.

Washington's response has been to zigzag. Rather than settling on a coherent framework, the administration has alternated between publicly supporting open models as an American strength and quietly pushing to restrict their distribution — leaving companies in Silicon Valley unable to plan around any consistent policy direction.

The whiplash raises a real question: can the U.S. champion open AI development while simultaneously trying to keep that technology out of adversaries' hands? The two goals are, at minimum, in tension, and the current approach of oscillating between them satisfies neither side of the debate.

Read the full story at NY Times Tech