The debate over who should control the most powerful AI systems is becoming harder to separate from the debate over who should be allowed to use them.
Mark Zuckerberg and Sam Altman have both argued that advanced AI should benefit people broadly, but their proposals put different weight on openness, safety standards, and government oversight.
On the one hand, Zuckerberg laid out his position in a 6,500-word manifesto published by Meta on August 10. Titled “The Future is for Everyone,” he argued that the biggest danger is the concentration of AI power in the hands of a single company, government, or other entities. His answer is to put increasingly capable AI tools into the hands of as many people as possible.
On the other hand, Altman has also warned about concentrated control, but he has paired that concern with a stronger case for international safety standards before the most advanced systems are widely distributed.
Zuckerberg Wants AI Power Spread Widely
Zuckerberg’s argument is closely tied to Meta’s strategy of releasing open-weight models. In his manifesto, he said everyone should eventually have access to free or affordable AI tools and argued that widespread access could create checks and balances against excessive control by powerful institutions.
Meta also announced the release of the weights for Muse Glimmer and said it would release the weights for a version of Muse Spark 1.2, with Zuckerberg saying Meta’s independent board would have authority to approve safety criteria for future model releases.
His position also goes beyond model access. Zuckerberg wants the U.S. and other democratic countries to maintain leadership in advanced AI and has warned that policies slowing American model releases could give foreign competitors an advantage.
Altman Wants Access With Rules Around It
Altman’s position is more complicated than simply advocating restricted access.
In a Financial Times article, he proposed a U.S.-led international forum that would establish common safety standards, independently assess AI capabilities and risks, and make advanced technology available to countries and companies that participate in the system as well as follow its rules. He argued that safety standards should come before broad distribution because powerful AI could cause significant harm if misused.
However, Altman argues the same as Zuckerberg, as he has repeatedly warned about the risks of AI becoming concentrated in the hands of a few companies or countries. In a July podcast appearance, he described a future where a small number of powerful actors control AI as a form of “AI authoritarianism” and argued for broader public access to the technology.
Two Visions for Who Gets to Shape AI
The significance of Zuckerberg and Altman’s positions goes beyond the companies they lead. Their arguments point to a growing question around advanced AI, which is whether its future should be shaped mainly by the companies building it, governments setting the rules, or the people using it.
Zuckerberg’s answer is clear in the sense that he, and at large, Meta, believes the safest path is to put increasingly powerful AI in the hands of individuals rather than allow a small group of institutions to control it. And Meta’s decision to release open-weight models alongside his manifesto puts that philosophy into practice.
Altman and OpenAI’s answer puts more weight on the role of institutions in managing the risks that come with that access.
Both approaches start from the idea that advanced AI should benefit more than a small group of companies or governments. Where they, however, differ is in how much control should remain with the people using the technology and how much should be exercised through the institutions responsible for managing its risks.
This debate is likely to become more important as AI moves from a tool controlled mainly by technology companies to a tech with much greater influence over how people work, create, and make decisions.
