Mark Zuckerberg Advocates for Decentralized AI Superintelligence

Meta CEO Mark Zuckerberg is championing a decentralized approach to artificial intelligence, arguing that superintelligence should be distributed to individuals rather than concentrated within a few powerful institutions. In a New York Times interview and a Wall Street Journal op-ed, Zuckerberg asserted that broad access to AI will empower people to improve health and start businesses, while preventing a handful of organizations from wielding undue influence over global politics and economics.

Zuckerberg has urged the federal government to accelerate AI development and remove regulatory bottlenecks, specifically criticizing a proposed 30-day review period for new models. He argues that the U.S. should focus on building superior systems to maintain global leadership rather than attempting to ban open-source models from China. To support this vision, Meta plans to invest between $125 billion and $145 billion in infrastructure in 2026 to integrate AI capabilities across its platforms.

This open-weight strategy contrasts sharply with the positions of rivals like OpenAI and Anthropic, whose leaders have highlighted alignment risks and the need to restrict powerful chips from authoritarian regimes. Furthermore, 1,178 employees from major labs including Google DeepMind and Microsoft signed an open letter titled "Pacing the Frontier," requesting that the government implement mechanisms to deliberately slow automated AI development to ensure human oversight. While Zuckerberg acknowledges the need for government coordination on biological risks, he maintains that an optimistic, open-source approach is the safest path forward.