Is Decentralized AI Governable? A Paradigm Shift from Policy to Protocol
The rise of decentralized artificial intelligence (DeAI) marks a transformative shift in AI development by leveraging distributed ledger technologies to decentralize data, computation, coordination, and economic models. Advocates highlight DeAI's potential to address key limitations of centralized AI systems—including single points of failure, trust deficits, and value monopolization. However, DeAI introduces significant challenges, particularly in governance. Unlike centralized systems, DeAI operates across global, borderless networks, making it resistant to traditional regulatory approaches. This paper investigates two central questions: (1) How do decentralized technologies enable "permissionless computation" as substrates for decentralized AI, leading to self-sovereign autonomy and potential unstoppability? (2) What regulatory and governance challenges does decentralized AI pose on permissionless computation substrates, and how can these be addressed? We identify the inherent characteristics of permissionless computation that make DeAI resistant to traditional oversight—including global reach, immutability, and adaptive survival strategies—rendering the technology "ungovernable" in the conventional sense. We contribute to raising awareness of DeAI governance research in this rapidly emerging field, emphasizing its potentially profound impact. Finally, we argue for a paradigm shift from traditional regulatory approaches toward technological protocol-based governance that integrates safeguards and adopts dynamic, emergent protocol designs to foster human-AI symbiosis.