Meta Platforms CEO Mark Zuckerberg has urged the United States to reduce restrictions on open-source AI to enhance competitiveness against Chinese counterparts. This call comes as the social media giant unveiled a new, smaller, and customizable AI model named Muse Glimmer, designed to run tasks on personal computers with a single graphics card. Zuckerberg hinted at the imminent release of even larger models in a video post accompanying a 14-page essay advocating for wider AI distribution.
The move towards open-weight AI has gained momentum as businesses seek cost-effective alternatives and raise concerns about cybersecurity incidents involving models from various companies. Open-weight models allow users to customize core components for specific purposes, contrasting with closed models that remain under a company’s control.
High-profile Chinese open-weight models like Kimi K3 and recent cybersecurity events involving AI models have underscored the importance of open-source AI. Meta, alongside other tech companies, signed a letter endorsing open-weight AI. Chinese startups have taken the lead in developing open-weight models, challenging U.S. developers whose models are predominantly closed.
Meta’s stocks rose nearly 3% in premarket trading following the announcement. The company plans to release Muse Spark 1.2, its most advanced model, to regain momentum in the AI race. Zuckerberg highlighted the need for the U.S. to revisit policies for open-weight models, emphasizing the importance of data use and distillation methods.
In a bid to allay concerns about its data-center expansion, Meta announced a $1 billion fund aimed at supporting communities impacted by its infrastructure projects. Zuckerberg stressed the need for the U.S. to streamline infrastructure development to compete effectively with countries like China. Meta intends to implement a governance structure to ensure safety criteria for model releases are approved by independent directors.
The U.S. faces challenges in fostering open-source AI due to stricter regulations compared to foreign counterparts. Zuckerberg emphasized the need to reduce friction in the development and adoption of American open-source models. The Trump administration recently indicated that open-weight AI models would not undergo voluntary safety tests, according to sources familiar with discussions.