Hugging Face, a leading artificial intelligence (AI) platform, has called on the US government to prioritize open-source AI development in its upcoming AI Action Plan. In a formal statement to the Office of Science and Technology Policy (OSTP), the company emphasized how open-source AI can fuel innovation, enhance security, and ensure the US maintains global leadership in AI advancements.
With over 1.5 million public models and seven million users across sectors, Hugging Face is at the forefront of the AI open-source movement. The company believes that a national focus on open AI ecosystems will accelerate technological progress, support economic growth, and democratize AI access.
Three Pillars of Hugging Face’s Proposal
Hugging Face’s recommendations for the AI Action Plan are built around three core pillars:
-
Strengthening Open-Source AI Ecosystems
- Hugging Face advocates for bolstered infrastructure support through initiatives like the National AI Research Resource (NAIRR).
- Investment in public data repositories and open science platforms will allow smaller organizations and academic institutions to contribute to AI innovation.
- The company underscores that breakthroughs in AI are often fueled by collaborative research and contributions from diverse stakeholders.
-
Facilitating Efficient and Reliable AI Adoption
- Hugging Face emphasizes the need for research and infrastructure investments that lower barriers to AI adoption.
- Promoting modular AI models and developing application-specific solutions can accelerate the spread of AI benefits across industries, including healthcare, finance, and manufacturing.
- By ensuring broader access to AI technologies, the US can foster a more inclusive AI economy.
-
Promoting Security and Standards
- Drawing from decades of experience in cybersecurity and information security, Hugging Face proposes adopting transparent security standards for AI systems.
- The company recommends implementing measures like traceability, disclosure requirements, and interoperability standards to ensure AI solutions remain secure and trustworthy.
Why Open-Source AI is Critical for US Leadership
Hugging Face argues that the AI landscape has evolved significantly due to open research. The company points to recent innovations like OLMO-2 and Olympic-Coder as examples of how open-source AI rivals, and sometimes surpasses, commercial models in efficiency and domain-specific performance.
According to Hugging Face, the rapid advancement of AI has resulted in more efficient models. While billion-parameter models were once necessary for complex tasks, similar results are now achievable with 2 billion parameter models. This compression of development timelines showcases the potential of open-source innovation.
Economic Impact
- Research indicates that open technical systems have a 2000x economic multiplier effect.
- A $4 billion investment in open-source AI could generate up to $8 trillion in value across industries.
- Countries without open-source AI contributions could face an average 2.2% loss in GDP.
- In 2018, open-source AI added between €65 billion and €95 billion to the European GDP.
These statistics underscore how prioritizing open-source AI can enhance the US economy while fostering innovation and competition.
Commercial Adoption: Practical Benefits of Open-Source AI
Hugging Face highlights the growing adoption of open-source AI in the private sector. Companies across industries are turning to open-source models for their cost efficiency, flexibility, and reduced reliance on proprietary systems.
Key Drivers for Open-Source Adoption:
- Cost Efficiency: Reduces research and development (R&D) expenses by using pre-existing open-source models.
- Customization: Enables organizations to tailor AI models to specific business use cases.
- Vendor Independence: Minimizes vendor lock-in, allowing companies to retain control over their AI deployments.
- Competitive Performance: Many open-source models now match or outperform proprietary AI systems.
Startups, mid-sized companies, and enterprises in sectors like finance and pharma increasingly rely on open models to stay competitive.
Hugging Face’s Policy Recommendations for Open AI Support
To create a thriving open-source AI ecosystem, Hugging Face proposes the following policy measures:
-
Enhance Research Infrastructure
- Fully implement and expand the NAIRR pilot to provide researchers with equitable access to computing resources, datasets, and collaborative tools.
-
Allocate Public Computing Resources
- Dedicate a portion of publicly funded computing infrastructure to support open-source AI projects, lowering the financial barriers for smaller organizations.
-
Enable Access to Data
- Establish policies that sustain public data ecosystems.
- Support the development of public data repositories and address challenges in accessing high-quality datasets.
-
Promote Rights-Respecting Data Access
- Develop clear guidelines for responsible data use, including standard protocols for data anonymization and consent management.
- Foster public-private partnerships to create secure data trusts.
-
Invest in Open Data Initiatives
- Support projects like the IBM AI Alliance Trusted Data Catalog and initiatives focused on digitizing public data repositories.
-
Foster Sector-Specific Innovation
- Provide resources and incentives for sector-specific AI adoption across industries like healthcare, education, and energy.
-
Expand AI Risk Management Frameworks
- Strengthen institutions like NIST to facilitate cross-sector collaboration and develop best practices in AI safety and risk management.
-
Support AI Model Evaluation
- Ensure access to robust and transparent datasets for evaluating AI model performance and reliability.
The Path Forward
Hugging Face’s recommendations align with the Biden administration’s efforts to establish ethical AI governance. The company believes that implementing an open-source-focused AI Action Plan will:
- Enhance the US’s global AI competitiveness.
- Democratize AI development and adoption.
- Ensure secure, transparent, and responsible AI deployment.
- Drive economic growth through collaborative innovation.
By prioritizing open-source AI, the US can strengthen its technological leadership while fostering an inclusive AI economy that benefits businesses, researchers, and society.