Mark Zuckerberg, CEO of Meta, has expressed concerns about the concentration of artificial intelligence (AI) technology and power in the hands of a few companies, declaring a commitment to expanding open AI. Meta has reentered the open AI competition by unveiling a new AI model that can run on personal computers.
On August 10, Zuckerberg published a 6,500-word essay titled "The Future is for Everyone," stating, "The idea that AI is so dangerous that we must concentrate power to ensure safety is fundamentally problematic."
Zuckerberg argued that superintelligence should be accessible and controllable by everyone rather than monopolized by a select few companies. He warned that concentrating AI technology within a few firms could stifle market competition and innovation.
On the same day, Meta introduced the 'Muse Glimmer,' an AI model featuring 30 billion parameters. This model is designed to operate with just one graphics processing unit (GPU) on standard PCs or Macs, allowing users to perform tasks such as coding and operating AI agents without an internet connection.
Meta also plans to release the weights of its top AI model, 'Muse Spark 1.2,' in the near future. Following underwhelming performance from its Llama 4 model, Meta has shifted back to an open strategy after previously focusing on closed model development.
Zuckerberg emphasized, "We will enhance safety measures alongside the expansion of openness." He stated that the safety standards and public availability of the new AI model will be reviewed by independent directors on Meta's board. He also proposed that governments collaborate with companies from the early stages of model development to assess potential misuse in cyberattacks or bioweapons development.
To support communities where AI data centers are established, Meta announced the creation of a $1 billion fund titled "The Future is for Everyone."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.