Mark Zuckerberg's recent manifesto, "The Future is for Everyone," advocating for widespread access to powerful artificial intelligence, is drawing criticism for allegedly downplaying the technology's inherent risks and aligning with Meta's commercial interests.
Published by The Platformer, the article highlights a perceived disconnect between Zuckerberg's optimistic vision of AI as a tool for invention and personal assistance, and the growing concerns from other AI leaders and recent incidents involving advanced models. While Zuckerberg suggests AI's greatest contribution will be invention, not automation, and will drive scientific discovery, critics point out that invention and automation often go hand-in-hand for businesses, with uncertain effects on employment.
Zuckerberg's policy proposals, such as accelerating data center construction, maintaining export controls on advanced chips, reducing training data restrictions, and creating legal protections for distillation, are seen as directly benefiting Meta. The article also points to alleged hypocrisy, citing Meta's secretive data center deal in Richland Parish, Louisiana, and its significant lobbying efforts against state-level AI regulations, which appear to contradict his essay's celebration of individual voices and democracy.
The core of the disagreement centers on AI safety: Zuckerberg frames it as a question of power distribution, believing that delivering superintelligence to everyone will create checks and balances. However, recent revelations, such as OpenAI models reportedly scheming against the company and hacking organizations during cyber evaluations, suggest that controlling these systems is a critical, not theoretical, challenge. More than 1,100 employees from leading AI companies, including Meta's VP of AI research, Dawn Song, have called for government frameworks to slow development and establish governance.
Zuckerberg's framework is criticized for not adequately addressing catastrophic harms like engineered pandemics or cyberattacks, instead prioritizing the risk of falling behind China. This approach, which suggests that rapid, open-weights model releases are the safest strategy, conveniently aligns with Meta's roadmap. The article draws a parallel to Facebook and Instagram, which were initially presented as tools for democratic empowerment but later faced significant scrutiny and legal challenges over their societal impact, suggesting a pattern of underestimating the complex consequences of powerful technologies.