Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
America since 2022 He had a strong lead in artificial intelligence This is thanks to advanced models from leading companies such as OpenAI, Google DeepMind, Anthropic, and xAI. However, a growing number of experts are concerned that the United States is starting to fall behind when it comes to creating open-weight AI models that can be downloaded, adapted, and run domestically.
Open models from Chinese companies such as Kimi, Z.ai, and Alibabaand Deep Sick These technologies are now rapidly gaining popularity among researchers and engineers around the world, leaving the United States behind in an increasingly important area of artificial intelligence innovation. “The United States needs open models to advance its leadership at every level of AI,” says Nathan Lambert, founder of the AI Foundation. ATOM (American Truly Open Models) project.“says Wired.
The most advanced models from US companies can only be accessed through a chatbot interface or by sending queries to companies’ servers through an application programming interface, or API. OpenAI and Google have released models with open weight, but they are much less capable than Chinese offerings, are more suitable for modification and offer more support for developers. Chinese model makers also benefit from open sourcing for their models, where the best ideas and modifications from outside researchers can be incorporated into future versions.
Lambert, who is also a researcher at the Allen Institute for Artificial Intelligence (Ai2), a non-profit in Seattle, Washington, founded the ATOM project to highlight the risks associated with the United States’ lag in open source. He says the country needs cutting-edge open models, in part because relying on foreign models could be a problem if those models are suddenly discontinued or become closed source.
Open models also foster innovation and experimentation among startups and researchers, Lambert says. Beyond that, companies with Sensitive information They need open models that they can run on their own hardware. “Open models are an essential part of AI research, deployment, and innovation, and the United States must play an active role in leading rather than following other contributors,” Lambert says.
Project Atom, launched on July 4, makes a compelling case for greater openness and shows how Chinese open-weight models have outperformed American models in recent years.
Ironically, the open source AI movement was launched by the American social media giant Meta, when it launched Release the llamaan open-weight parametric model, in July 2023. At the time, Meta saw Llamas as a way to break into the AI race. Her new model quickly became popular among researchers and entrepreneurs.
Since then, Meta and other US AI companies have become fixated on the idea of developing human-level or superhuman-level AI, preferably before their competitors, resulting in less openness. In recent months, Zuckerberg has relaunched Meta’s AI efforts with… A series of expensive appointments And the new “super intelligence” laboratory. Zuckerberg He also pointed out That Meta no longer open source their best models.
In contrast, China’s technology industry has moved toward greater openness this year. In January 2025, DeepSeek, a then little-known startup, released software An open model called DeepSeek-R1 Which shook the world due to its advanced abilities and the fact that it trained at a fraction of the cost of the top American models. Since then, a number of Chinese companies have introduced powerful open-weight models featuring additional innovations.