Zuckerberg: "High-Performance AI Should Not Be Monopolized by a Few Companies or Governments"
Mark Zuckerberg, CEO of US big tech company Meta, has argued that high-performance artificial intelligence (AI) capabilities should not be concentrated in the hands of a few companies or governments. He explained that empowering individuals and striking a balance of power are fundamental to AI safety.
On August 10 (local time), Zuckerberg published a 6,500-word essay titled "The Future is for Everyone," in which he stated, "There is a fundamental problem with the idea that AI is so dangerous that only extreme centralization of power can guarantee safety." He emphasized that rather than allowing superintelligence to be concentrated among a select few, it should be accessible to everyone.
He presented his philosophy that individual empowerment is the source of prosperity and that a balance of power should be the foundation of AI safety. "When people are empowered, they naturally check and balance each other in all areas—economy, society, and politics—and in turn, institutions such as companies and governments are kept in check," Zuckerberg said. "The key to a future that benefits everyone is establishing a balance of power in favor of individuals," he added.
Zuckerberg also offered criticism seemingly directed at competitors like Google, Anthropic, and OpenAI. He asserted, "Most other AI labs are focused on developing AI for institutions such as corporations and governments," arguing that if these companies dominate the market, power will be further concentrated in large institutions rather than individuals.
Zuckerberg proposed that the US government and AI firms should establish a framework to jointly verify the safety of new AI models and that AI developers should be allowed to use "model distillation," a process where one model is used to improve another. He said, "AI labs should not wait until training is entirely complete on a new model, but should share intermediate checkpoints with the government for oversight and review," adding, "This would allow the government early access to the most powerful models and enable experienced engineers to identify and address security issues."
Zuckerberg also rejected the outlook that AI will cause large-scale job losses. He argued that if AI is used to enhance individual capabilities and to create new products and services, rather than simply automating existing work, entrepreneurship and employment could actually increase. "No matter how intelligent AI becomes, computational resources will always be limited, so there will be an opportunity cost depending on how they are used," he emphasized. "It may be more reasonable to use AI to invent something truly valuable rather than automating existing jobs." He added, "In the future, we may see a small number of people operating sizable companies with personal superintelligent agents."
He also stressed that the United States and its allies must lead the global open-source AI ecosystem. Given that US companies face more regulatory restrictions than foreign competitors, especially regarding the use of training data, he insisted that such constraints must be relaxed to enhance the competitiveness of US open-source models. "For the US to maintain leadership in open-source, policies must be reviewed across multiple domains, including model distillation and access to training data," he said. "I do not believe that simply restricting access to overseas open-source models is an effective solution."
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
Meanwhile, Meta unveiled its new open AI model "Muse Glimmer" on the same day. Muse Glimmer specializes in on-device AI. Meta also plans to release the weights for a higher-performing version, "Muse Spark," within the next few weeks. The Financial Times (FT) noted that this move reflects a return to Meta's "open AI" strategy, which the company has championed to distinguish itself from competitors.
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.