Two dozen technology companies and organizations, including Nvidia, Meta, IBM, and Microsoft, issued a joint letter on Friday urging U.S. lawmakers to support the development of open-source artificial intelligence. The group argued that open models promote competition and allow for broader security auditing by the research community. Nvidia CEO Jensen Huang stated in the letter that lawmakers should avoid what he described as premature restrictions that could impact innovation and international competitiveness.
The appeal comes as members of Congress consider legislation to regulate AI following reports of a cyberattack involving an OpenAI model. Proposed measures include a requirement for "kill switches" that would allow for the immediate deactivation of AI models. Additionally, the Trump administration has reportedly considered sanctions against Chinese developers of open-source AI models, citing concerns over the alleged theft of proprietary U.S. technology.
Advocates for open-source AI contend that making model weights public allows developers to identify vulnerabilities and build safeguards. They also noted that restricted, closed-source models are not immune to breaches. Conversely, proponents of tighter regulations and closed-source models argue that unrestricted access to powerful AI software poses safety and national security risks. Some tech executives, including those from Microsoft and Palantir, have also cited lower infrastructure costs as a primary reason for their support of open-source initiatives.