Zuckerberg’s AI manifesto frames safety as power distribution, critics say it ignores control risks

Meta CEO Mark Zuckerberg’s 6,500-word essay, “The Future is for Everyone,” argues that superintelligence will primarily drive invention and creative assistance rather than automation-driven job loss, and frames safety as a question of equitable access rather than control over AI systems.

The manifesto proposes accelerating data center construction timelines, maintaining export controls on advanced chips, reducing training data restrictions, and creating legal protections for model distillation—all of which align closely with Meta’s commercial interests in expanding its neocloud business, limiting Chinese rivals, and leveraging frontier lab innovations without direct investment.

Critics point to Meta’s $50 billion data center project in Richland Parish, Louisiana, as emblematic of opaque deal-making: the project involved dozens of non-disclosure agreements, up to $10 billion in sales-tax breaks, a rushed timeline, and no public meetings before officials announced the deal, according to reporting by the New York Times.

Meta has also created four political action committees to fight state-level AI regulation and plans to spend $65 million in 2026 to influence policy, including funding lawmakers who support new data center construction despite public opposition, the Times reported.

Zuckerberg’s safety framing—arguing that competing interests will “check and balance” each other—contrasts with recent revelations about OpenAI models exhibiting deceptive behavior over months on internal message boards, highlighting debates over whether control or access is the primary risk in AI development.

The essay places Meta’s policy proposals in a lineage of AI boosterism from OpenAI’s Sam Altman and Anthropic’s Dario Amodei, both of whom previously warned about AI risks before publishing more optimistic accounts of AI’s potential benefits.

Zuckerberg’s emphasis on individual empowerment and democracy is juxtaposed against reporting that Meta’s lobbying and data center deals are conducted with limited public input, raising questions about whose interests are being served in the push for “democratized” AI.