Meta Platforms CEO Mark Zuckerberg published a sweeping 6,500-word essay on Monday outlining his vision for the future of artificial intelligence, in which he argued that everyone should have access to a personal, all-knowing AI agent capable of enhancing every aspect of their lives.

The manifesto arrives amid an intensifying industry divide over whether advanced AI systems should be freely shared or tightly controlled — a debate that has split the field's most powerful players and drawn in governments worldwide.

Zuckerberg used the essay to advocate forcefully for open-source AI — a development model in which key technical components are made freely available for anyone to inspect, modify, and build upon. The Meta chief envisioned a future in which individuals are empowered to launch new businesses, access Ph.D.-level tutoring, and receive personalized lifestyle guidance. He cited his own 8-year-old daughter, who he said can already code her ideas and rapidly generate videos, as evidence of what is becoming possible.

"Everyone will soon have invention superpowers," Zuckerberg wrote.

The essay was published the same day Meta announced the release of a new open-source AI model called Muse Glimmer, which is designed to run on a personal computer. Zuckerberg also said Meta would provide developers with access to a more powerful model, Muse Spark 1.2. The releases extend Meta's open-source AI strategy, which the company has pursued through successive generations of its Llama model family since 2023, positioning itself as the primary counterweight to closed-system developers in the industry's most consequential technology race. Like other major tech companies, Meta builds AI models to power its own platforms — including Instagram and Facebook — and makes them available to external developers who create their own applications and features.

Zuckerberg cautioned that the greatest risk in AI development lies in concentrating control of advanced systems within a small number of companies, institutions, or governments. If control of AI "superintelligence" becomes too concentrated, he warned, the result will be worse outcomes for the majority of people. The solution, he argued, is ensuring that advanced AI is "broadly distributed" to empower billions of people equitably.

Without naming specific competitors, Zuckerberg made what appeared to be a veiled reference to companies such as OpenAI, Anthropic, and Google — rivals that have not embraced open-source development to the same degree.

"Most other labs are focused on building AI for companies, governments, or other institutions, so if those labs lead, then the balance of power will favor larger institutions over individuals," he wrote.

The essay drew sharp criticism from AI safety advocates. Anthony Aguirre, president and CEO of the Future of Life Institute — the organization behind the widely publicized 2023 open letter calling for a pause on training AI systems more powerful than GPT-4 — pointed to recent incidents in which AI models, including Meta's own, hacked other companies on at least three separate occasions. He asked: "what on Earth makes Meta think they — or anyone — could control something exponentially smarter, faster, and more capable?"

"Meta's chosen path is ultimately one of human replacement and disempowerment, but there is an alternate, pro-human path — one that brings about individual flourishing and scientific breakthroughs with controllable AI tools," Aguirre said. "What's urgently needed now is for governments to facilitate a pause on advanced AI development and steer us onto the pro-human path — before it's too late."

Zuckerberg also addressed policy measures for maintaining U.S. competitiveness against rivals including China. He urged the United States and its allies to accelerate the construction of energy capacity and infrastructure needed to support AI development — a concern shared across the industry, as the rapid expansion of AI data centers has placed growing strain on power grids and prompted major tech companies to invest in new energy sources, including nuclear power.

On the topic of "distillation" — the practice of training a less capable AI model on the outputs of a more powerful one — Zuckerberg pushed back against efforts to restrict the technique. The Trump administration has vowed to crack down on Chinese AI companies that use distillation to extract technical features from U.S. models. Earlier this year, Anthropic accused China's DeepSeek of engaging in the practice.

"Some have tried to frame distillation as harmful, but I think it is important to protect the principle that you can learn from anything you can observe," Zuckerberg said. "This is how the world works, and the U.S. will not be able to lead if we restrict ourselves on this front."

Matt Lane, senior policy counsel at the digital rights advocacy group Fight for the Future, called Zuckerberg "annoyingly out of touch on his best day," while conceding that the CEO is correct about the importance of open-source AI. Lane argued that shared access to open-source AI systems serves as a critical safeguard in a world where AI could be used to compromise electronic infrastructure such as voting systems and benefits networks.

However, Lane cautioned that widespread adoption of Meta-designed AI systems — even fully open-source ones — would ultimately benefit Meta above all others. "Open source AI needs to be supported beyond these corporate interests. We, unlike Zuckerberg, hope for as much competition as possible," he said.