Mark Zuckerberg Warns Against Centralized AI Power

Meta CEO Mark Zuckerberg argued in a Wall Street Journal op-ed that the development of artificial intelligence should not be controlled by a small group of major institutions. Zuckerberg advocated for the creation of personal superintelligence available to all users to prevent the centralization of power over technology that could eventually surpass human intelligence, according to The Hill.

Zuckerberg focused his critique on the prospect of superintelligence and the distribution of its access. In the op-ed, he wrote that the primary question of the current era is whether such technology will be restricted to a few institutions or serve as a tool that empowers everyone, as reported by The Hill.

The Meta chief executive specifically challenged the perspectives of AI rivals, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei. Zuckerberg expressed surprise that leaders in the field have offered views filled with doom, according to The Hill.

Zuckerberg questioned the logic of building a future that some believe will eliminate most jobs and human relevance. This follows warnings from Amodei and Altman regarding AI-related job losses. The Hill reports that Amodei suggested last year that AI could eliminate significant portions of entry-level white-collar work, while Altman has previously warned of job losses, though he later walked back some of those predictions.

Zuckerberg rejected the idea that extreme concentration of power is the only safe path to manage dangerous AI. He stated that historically, trusting absolute power to benevolently provide for humanity has not led to positive outcomes, according to The Hill.

Meta’s Strategy for Open-Weight Models

Meta has integrated this vision of decentralized AI into its corporate strategy. Zuckerberg previously detailed his goals for personal superintelligence in a letter written last July, according to The Hill.

The push for open access comes as the American AI sector faces new competition from China. The Hill reports that the emergence of open-source or open-weight models from Chinese firms, such as Moonshot AI’s Kimi K3, has disrupted U.S. companies that have primarily relied on closed models.

The Hill reports these developments have raised concerns that U.S. firms may shift toward cheaper, open models produced in China.

In response to these market shifts, several major AI firms have lobbied the U.S. government. According to The Hill, Meta, OpenAI, and Nvidia signed a letter last week urging the White House not to overregulate open-weight models.

Democratic Access vs. Centralized Control

Zuckerberg argued that there is no single objective answer to how people define the best life, citing the history of democracy and economics. He stated that the best approach is allowing individuals to decide what matters to them rather than letting a small set of experts decide what is best for humanity, according to The Hill.

Rather than centralizing this power, we believe that delivering personal superintelligence to everyone is the way to answer this question. — Mark Zuckerberg

This position places Meta in direct opposition to the closed-model approach favored by several of its primary competitors. While Meta pushes for widespread access to the weights of its models, other leading U.S. firms have maintained tighter control over their proprietary systems, as reported by The Hill.