Nvidia is arguing limiting access to open-source artificial models could concentrate power in the hands of a few, large artificial intelligence (AI) companies.
Nvidia on Monday launched the Open Secure AI Alliance, which includes many leading tech companies such as Palantir, IBM, Crowdstrike and SpaceX. The chipmaker maintains a predominant control of the AI hardware market, as Bloomberg found that it will control 81% of the AI accelerator hardware market, or chips that AI platforms use for machine learning, by 2033.
The Trump administration is considering limiting access to Chinese open-source models, Axios reported. The Treasury Department could sanction Chinese AI labs for alleged intellectual property theft, Secretary Scott Bessent said. The Trump administration now leans towards limiting access to open-source models after previously taking a hands-off approach to regulating AI, the outlet reported.
The administration’s renewed interest in regulating open-source models has coincided with the release of Chinese AI company Moonshot AI’s latest model, Kimi K3. Moonshot experienced so much demand after its release that it paused new subscriptions.
“Blanket restrictions on open frontier AI systems would weaken defensive capacity and risk concentrating power, dependence and vulnerability in a few closed providers,” Nvidia stated in its press release, urging regulators and policymakers to realize the value of open-source models.
“Companies and governments should invest in shared open infrastructure for AI defense — datasets, evaluation frameworks, attack simulators and red-teaming tools — much as past generations invested in open source software,” the company added. The company said that protecting open-source AI could “strengthen competition.”
An Nvidia spokesperson referred the Daily Caller News Foundation to company’s press release announcing the coalition.
“Attackers have frontier AI. Defenders need a frontier AI ecosystem—the best open and closed models, force-multiplied by a global community,” Jensen Huang, the founder and chief executive of Nvidia, wrote on Monday.
Hackers and criminals can exploit open-source AI models because they do not have guardrails set by closed-source AI platforms such as Anthropic and OpenAI, according to research conducted by SentinelOne and Censys, two cybersecurity companies. Bad actors can use open-source AI for scams, fraud and hacking.
“Illicit, adversarial distillation is IP theft and industrial espionage that supports adversary military and intelligence capabilities. It is a national challenge that creates serious national security risks for the United States and democratic allies,” Sarah Heck, the head of public policy for Anthropic, wrote in mid-July about how Chinese AI models use American AI to boost Chinese AI models through a process known as distillation. Anthropic said distillation can help AI develop models at far less cost compared than it would normally take to develop a model.
“The President’s team is working closely together to balance innovation and security in AI policymaking,” White House spokesperson Liz Huston said in a statement to Politico on Thursday.
(DCNF)