Zuckerberg: the United States and its allies should lead open-source AI

Mark Zuckerberg is calling for «the United States and its allies to lead the open-source artificial intelligence ecosystem». The line appears in an open letter titled «The Future is for Everyone: The Path to a Positive AI Future».

On the same day Meta released two new models, Muse Glimmer and Muse Spark 1.2, both open source. It is a strategic choice that mirrors the approach already taken by Chinese competitors.

The letter's central argument concerns market structure. For Zuckerberg «open source is a positive and important force for empowering people and preventing centralisation, which is harmful both to security and to the economy». Concentration is thus presented not only as a technical risk but as an economic problem.

Looking ahead, the head of Meta writes about superintelligence. He describes it as «one of the most important technologies in history», with opportunities and challenges potentially greater than anything experienced so far.

Three principles

To guide that development the letter puts forward three principles. The first sees individual empowerment as the source of prosperity. The second identifies invention as the main purpose of superintelligence. The third makes the balance of power the foundation of security. Distributing superintelligence widely, Zuckerberg argues, is the key to empowering individuals.

The text leaves several questions open. It does not say who is to bear the cost of models given away for free, nor how a call for American leadership fits an ecosystem that by definition anyone can use, copy and modify. The effects on competition are also unclear: opening your own models lowers the barriers for newcomers, yet the same move can entrench the position of whoever produces those models and sets the standard.