The AI company, Hugging Face, has submitted its recommendations to the USA government regarding how it should treat AI and how the government can use AI for the country’s own advantage. The rise of AI in recent years has given birth to differences among different US governments and tech giants, with Hugging Face favoring open-source and collaborative development of AI as the best strategy to fulfill White House AI Action Plan.
As per the recommendations given by Hugging Face, open-source AI models are better than closed commercial systems as demonstrated by OlympicCoder, which outperforms Claude 3.7 using only 7 billion parameters, and open OLMo 2 models.
It suggests that the federal government should collaborate with the private sector, which would be productive for AI development in the country. But this is totally opposite to the stance taken by one of the commercial AI leaders, OpenAI, which advocates for the minimal involvement of the state regulations and voluntary collaboration between the state and the private sector. Not only this, but both AI leaders also differ about the type of datasets to be used to train AI models. Hugging Face favours trusted datasets only, but OpenAI has supported the use of copyrighted material to train AI models.
Following Executive Order 14179, issued in January, Hugging Face has proposed a three-pillar plan to develop AI in the USA. According to its first pillar, open-source systems should be promoted. The company argues that open-source models not only will give advantage to the USA in the competitive world, but also will bring economic boom. It suggests that the country should establish open-source ecosystems by investing in National AI Research Resource ( NAIRR).
Its second pillar emphasizes the need to invest in specialized models that require limited resources and which smaller organizations can also afford. The country could widen the ambit of the AI ecosystem by allowing small organizations to participate in the AI development.
The third pillar focuses on security which has been a major issue for the US government. The company says that open and transparent AI systems would allow access to training data and procedures, thus making them more secure in critical applications than closed commercial AI models, which might be an oversimplification of the issue.
The venture capital firm, Andreessen Horowitz, has also taken a somewhat similar position regarding the role of the state in the development of AI. The company has supported the idea of federal leadership, but also favours strict regulations on the harmful aspects of AI.
Whether these recommendations by Hugging Face will be acceptable to the US government entirely depends upon the economic and security aspects of the open-source systems. If the company is able to show that open-source systems are secure from any Chinese threat, it might prevail in influencing the US-AI policy.
Image: DIW-Aigen
Read next:
• Hackers Bypass 2FA Using Stolen Session Cookies, SpyCloud Reports 17.3 Billion Stolen in 2024
As per the recommendations given by Hugging Face, open-source AI models are better than closed commercial systems as demonstrated by OlympicCoder, which outperforms Claude 3.7 using only 7 billion parameters, and open OLMo 2 models.
It suggests that the federal government should collaborate with the private sector, which would be productive for AI development in the country. But this is totally opposite to the stance taken by one of the commercial AI leaders, OpenAI, which advocates for the minimal involvement of the state regulations and voluntary collaboration between the state and the private sector. Not only this, but both AI leaders also differ about the type of datasets to be used to train AI models. Hugging Face favours trusted datasets only, but OpenAI has supported the use of copyrighted material to train AI models.
Following Executive Order 14179, issued in January, Hugging Face has proposed a three-pillar plan to develop AI in the USA. According to its first pillar, open-source systems should be promoted. The company argues that open-source models not only will give advantage to the USA in the competitive world, but also will bring economic boom. It suggests that the country should establish open-source ecosystems by investing in National AI Research Resource ( NAIRR).
Its second pillar emphasizes the need to invest in specialized models that require limited resources and which smaller organizations can also afford. The country could widen the ambit of the AI ecosystem by allowing small organizations to participate in the AI development.
The third pillar focuses on security which has been a major issue for the US government. The company says that open and transparent AI systems would allow access to training data and procedures, thus making them more secure in critical applications than closed commercial AI models, which might be an oversimplification of the issue.
The venture capital firm, Andreessen Horowitz, has also taken a somewhat similar position regarding the role of the state in the development of AI. The company has supported the idea of federal leadership, but also favours strict regulations on the harmful aspects of AI.
Whether these recommendations by Hugging Face will be acceptable to the US government entirely depends upon the economic and security aspects of the open-source systems. If the company is able to show that open-source systems are secure from any Chinese threat, it might prevail in influencing the US-AI policy.
Image: DIW-Aigen
Read next:
• Hackers Bypass 2FA Using Stolen Session Cookies, SpyCloud Reports 17.3 Billion Stolen in 2024
