Meta CEO Mark Zuckerberg advocates for open AI models in new manifesto

Mark Zuckerberg has released a manifesto outlining an aggressive vision for the future of artificial intelligence, emphasizing a push toward open models. He argues that AI models should be open, allowing users to download and modify weights, to prevent the concentration of power within a few closed systems like ChatGPT, Claude, or Gemini.

Critics suggest this stance may be a strategic marketing move to position Meta as an alternative to the AI safety movement, especially as the company seeks to catch up to competitors like OpenAI and Anthropic. While open models offer advantages in forensic analysis and code examination, some observers note that Meta has not yet opened its advertising recommendation algorithms, which maintain its market position alongside Google.

Zuckerberg’s vision includes the development of personal superintelligences that possess deep individual knowledge, supported by massive investments in data center infrastructure and the development of next-generation AI agents.