The White House is preparing to include open-source AI models in its regulatory framework for the first time, according to sources familiar with the matter. This shift could bring models like Meta's Llama 3.1 under federal oversight, marking a significant change in the administration's approach to AI governance.
TL;DR
- The White House is planning to regulate open-source AI models for the first time, potentially impacting models from Meta, Mistral, and other developers.
- This move comes as open-source models like Meta's Llama 3.1 and Mistral's offerings push the boundaries of AI capabilities.
- The regulatory update aims to address national security concerns while balancing the need for innovation and open development.
What happened
The White House is preparing to expand its AI regulatory framework to include open-source models, according to sources who spoke with WIRED. This marks a significant shift in the administration's approach to AI governance, as it has previously focused on closed commercial models from companies like OpenAI, Anthropic, and Google.
The timing of this update is critical, as Meta recently released Llama 3.1, its most capable open model to date, and Mistral continues to push boundaries in Europe. A growing ecosystem of startups is building businesses around freely available model weights, raising questions about the risks and benefits of open-source AI development.
The White House's planned update would need to thread a narrow needle, balancing the need for oversight with the desire to avoid stifling innovation. Industry insiders expect the framework to focus on transparency requirements rather than outright restrictions, such as mandatory disclosure when releasing models above certain capability thresholds or safety testing protocols.
Why it matters
This regulatory shift could have significant implications for AI developers, startups, and investors. For companies like Meta and Microsoft, which have invested in open models, the stakes are enormous. The next few months will determine whether open development remains viable at the frontier of AI capabilities or if regulatory pressure pushes the most advanced work back behind closed doors.
The update also brings US policy closer to the EU's AI Act, which attempts to regulate AI systems based on risk levels rather than development approach. Europe has struggled with how to handle open-source models within that framework, and Brussels will be watching closely to see how Washington tackles the same challenge.
For national security officials, the move addresses concerns about adversaries using open models to develop capabilities without the safety guardrails baked into commercial systems. However, the open-source community argues that transparency and distributed development actually make these models safer, with more eyes catching problems and faster fixes.
Key facts
- The White House is planning to include open-source AI models in its regulatory framework for the first time, according to sources familiar with the matter (WIRED).
- This shift could bring models from Meta, Mistral, and other open-source developers under federal oversight (WIRED).
- Meta recently released Llama 3.1, its most capable open model to date (WIRED).
- The regulatory update aims to address national security concerns while balancing the need for innovation and open development (WIRED).
- Industry insiders expect the framework to focus on transparency requirements rather than outright restrictions (WIRED).
- The update brings US policy closer to the EU's AI Act, which regulates AI systems based on risk levels (WIRED).
- The White House hasn't publicly commented on the planned update, and sources cautioned that details are still being finalized (WIRED).
- The administration has concluded that its current hands-off approach to open models isn't sustainable as capabilities continue advancing (WIRED).
Context
The White House's move to regulate open-source AI models marks a turning point in how governments approach AI governance. The administration is betting it can create meaningful oversight without crushing the innovation that makes open development valuable. However, the fundamental challenge remains: how do you regulate something designed to be ungovernable?
For companies betting their strategies on open models, the next policy draft can't come soon enough. The answer to this regulatory challenge will shape not just US AI policy but the global balance between open and closed development as model capabilities continue their rapid advance.
The current framework, built primarily around closed commercial models from OpenAI, Anthropic, and Google, doesn't capture the unique challenges of open-source development. When anyone can download model weights and fine-tune them for specialized tasks, traditional regulatory levers like API restrictions or usage monitoring don't apply. This has created a policy gap that's becoming harder to ignore.