AI Policy @🤗: Response to the White House AI Action Plan RFI

Source: HuggingFace Date: 2025-03-19 URL: https://huggingface.co/blog/ai-action-wh-2025

Summary

Policy statement: HuggingFace’s formal response to the White House Office of Science and Technology Policy’s AI Action Plan RFI. Three recommendations: (1) recognize open source/open science as foundational to AI, (2) prioritize efficiency and edge-deployable models, (3) secure AI through transparent/auditable open systems. Cites OlympicCoder-7B outperforming Claude 3.7 on coding tasks and OLMo 2 matching o1-mini with fully open training data as evidence.

Implications

Thread: open-weights ecosystem health / HF as open-source ML hub. HF is positioning itself as the open-source lobby for AI at the federal level — this is a strategic move, not just advocacy. The “open = more secure” framing (drawing parallels to open-source cybersecurity) is the most interesting policy argument: it directly counters the “dangerous open weights” narrative from closed-model incumbents. If the White House AI Action Plan adopts language favorable to open models, it could affect export controls, procurement rules, and research funding in ways that benefit HF’s ecosystem.