Meta CEO Mark Zuckerberg called on Monday for advanced artificial intelligence to be distributed widely, warning that concentrating superintelligence in the hands of a small number of companies, governments or institutions could pose significant risks.
In a letter titled The Future is for Everyone: The Path to a Positive AI Future, Zuckerberg argued that superintelligence should primarily be used to expand people’s ability to create, discover and solve problems, rather than simply automate existing human work.
He proposed an approach centred on individual empowerment, technological invention and a balance of power, arguing that these principles could help ensure advanced AI benefits society broadly.
Zuckerberg said no single superintelligent AI system could adequately represent humanity’s diverse and sometimes conflicting values and interests. Instead, he advocated distributing advanced AI capabilities among individuals, businesses and governments, with multiple AI systems providing checks on one another.
“The key to a positive future for everyone is achieving a balance of power that favors individuals,” Zuckerberg wrote.
The letter coincided with Meta’s launch of Muse Glimmer, an open-weight AI model containing 30 billion learned parameters and designed to operate AI agents locally on consumer devices.
Unlike AI assistants that depend heavily on cloud computing, Muse Glimmer can reportedly run on a Mac or PC equipped with a single consumer GPU, enabling it to carry out multi-step tasks without an internet connection.
Running the model locally could also allow users to keep more of their data on their own devices, rather than transmitting it to remote servers for processing.