OpenAI Governance Controversy: Thoughts on the DAO Model

Recently, a governance turmoil within OpenAI has attracted widespread attention in the industry. As the situation develops, more and more people are starting to ponder: If OpenAI were to adopt a DAO (Decentralized Autonomous Organization) model for governance, could this have prevented the turmoil?

As a non-profit organization dedicated to creating safe general artificial intelligence (AGI) that benefits all of humanity equally, OpenAI has, in some ways, become very similar to many DAO organizations that create public goods. However, the root of this turmoil does not stem from the organizational structure itself, but rather from the ambiguity and unreasonableness of its governance rules.

The main issues currently facing OpenAI include an insufficient number of board members and a lack of transparency in the decision-making process. For example, the original board of 9 members has now been reduced to 6, and key decisions (such as replacing the CEO) appear to have not been fully discussed and reviewed by the entire board. This practice not only lacks transparency but also fails to adequately consider the opinions of various stakeholders.

In comparison, the DAO model may offer some beneficial insights for OpenAI. For instance, introducing more checks and balances (such as employee representatives), and establishing a more transparent and inclusive governance mechanism. These practices can help OpenAI build a more robust, open, and inclusive governance structure.

It is worth noting that DAO and AGI have similar goals in pursuing autonomy. Both aim to build systems that can operate independently and are not controlled by external forces. However, in practice, how to balance autonomy with the necessary human intervention remains a question that requires in-depth discussion.

The latest developments show that about 90% of OpenAI employees are willing to follow Sam Altman in his departure. This phenomenon reflects a long-standing debate in the DAO field: in organizational governance, which is more important, the rules constrained by code or the consensus of the community?

Although rules and constraints can form many consensuses, true greatness in consensus often stems from a shared sense of mission and cultural values. This deep sense of identification can create strong resonance between people. However, when we turn our attention to AI, how to cultivate this resonance within artificial intelligence systems remains a topic worthy of in-depth exploration.