Zuckerberg warns AI's biggest risk is centralized control

Meta CEO Mark Zuckerberg has entered the artificial intelligence debate with a strong stance against concentrated power. In his "The Future is for Everyone" manifesto, Zuckerberg argues that the primary danger of AI lies not in its potential capabilities, but in the possibility of a single government or company controlling it.

“

The greatest risk of artificial intelligence is that a single entity or government has too much control over the technology.

He champions the broad distribution of superintelligence, asserting it will empower individuals and foster a more entrepreneurial economy. This philosophy is underscored by Meta's simultaneous release of Muse Glimmer, an open-source version of its powerful AI model, designed to compete with closed-source alternatives.

“

The notion that AI is so dangerous that the only safe path is extreme concentration of power is problematic.

Zuckerberg implicitly criticizes rivals like Anthropic and OpenAI, who have called for stringent AI controls. He warns that delaying the release of U.S. AI models could cede leadership to foreign regimes, specifically mentioning China. The Meta chief predicts that superintelligence will ultimately create more jobs, making it easier for people to start businesses without massive capital investment.

“

Invention, not automation, will be the greatest contribution of superintelligence.

To support his vision, Zuckerberg also announced "The Future is for Everyone Fund," a $1 billion initiative aimed at benefiting communities in the United States where Meta operates data centers. The fund will support local teachers, first responders, and infrastructure for energy and water.

“

Too much intervention or any policy that delays the release of U.S. models, even by a month, could add significant risk to U.S. leadership against foreign models, allowing regimes like China to take the lead.