By Stuart Kerr, Technology Correspondent
Published: 16 July 2025 | Last updated: 16 July 2025
Contact: liveaiwire@gmail.com | Follow @LiveAIWire
Author Bio: https://www.liveaiwire.com/p/to-liveaiwire-where-artificial.html
Intelligence Without a Master
Artificial intelligence thrives on centralisation. The largest models—trained by the likes of OpenAI, Google DeepMind, and Anthropic—are developed in fortified data centres by elite teams and distributed from the top down. But what happens when we turn this structure inside out?
Can intelligence be decentralised? And more importantly, should it be?
As AI grows more powerful, questions about control, access, and equity are intensifying. At the intersection of these debates lies blockchain technology—a system built for decentralisation, transparency, and trust without gatekeepers. Now, the fusion of AI and blockchain is beginning to reshape the very infrastructure of intelligence.
Code Meets Consensus
Blockchain, in its most basic form, is a decentralised ledger that records data immutably across a network. It’s the foundation of cryptocurrencies, smart contracts, and decentralised apps—or dApps. The Ethereum Foundation describes these tools as systems that operate autonomously, without central control.
This ethos appeals to AI researchers who fear monopolisation of knowledge. In Invisible Infrastructure, we explored how hidden systems influence power. Decentralised AI aims to do the opposite: to distribute computational control and give users a stake in how AI evolves.
But decentralisation is more than an ideology. It offers technical advantages too. Federated learning, for instance, enables AI models to be trained across multiple devices without pooling data centrally. Combined with blockchain, this approach offers unprecedented data privacy.
Trustless AI? Not Quite
Despite the hype, the merger of AI and blockchain is still fragile. Trustless systems—the ideal where no party needs to be trusted—collide awkwardly with machine learning, which inherently involves probabilistic inference and opaque parameters.
In practice, blockchain can audit and verify AI decisions. It can register provenance: where data came from, how models evolved, and who has rights to intervene. This is especially relevant in finance, where Code Capital explored how AI rewires digital trust. Smart contracts could one day govern when an AI is allowed to execute a transaction or halt a process.
The European Investment Bank’s report on AI and Blockchain frames this as an emerging regulatory challenge. If decentralised AI becomes a global infrastructure, it will need protocols that support both transparency and adaptability.
The DAO Brain
One of the most radical visions comes from the world of DAOs—Decentralised Autonomous Organisations. These are blockchain-governed entities that make decisions collectively, often without human CEOs or boards. What if an AI could serve as a DAO’s decision-making core?
The Medium article on AI in Web3 outlines this futuristic vision: AI agents governed by token-holder votes, continuously evolving based on user input. In theory, such systems could coordinate everything from urban planning to disaster relief.
But this raises sharp questions. Who trains the model? Who audits the outcome? And what happens when decentralised consensus leads to harmful or irrational behaviour?
In The AI and the New Feudal Web, we warned against algorithmic overlords. DAOs risk falling into similar traps—especially when AI replaces human deliberation with automated outputs.
Edge AI and Swarm Intelligence
Where decentralised AI gets especially interesting is at the edge—literally. In low-power devices, sensors, and phones, models can operate independently and collaborate like a digital swarm. The ArXiv paper on federated learning and blockchain presents early frameworks for these systems, balancing efficiency and accountability.
Rather than relying on one massive model, decentralised AI envisions thousands of micro-models that share insights via encrypted ledgers. No single failure collapses the network. No single actor controls the truth.
This vision isn’t just theoretical. MIT Media Lab’s decentralised AI project is experimenting with human-AI collaborations that decentralise not only processing, but decision-making itself.
A Fork in the Brain
Decentralising intelligence is seductive. It offers a pathway away from concentrated power, corporate surveillance, and algorithmic monopolies. But it also risks fragmentation, incoherence, and new forms of manipulation.
To decentralise the brain is to rethink what cognition means in a digital society. It forces us to ask: can we trust a mind with no master? Can distributed agents form coherent decisions? And if they can, do we still call it artificial intelligence—or something else entirely?
About the Author
Stuart Kerr is the Technology Correspondent at LiveAIWire. He writes about AI’s impact on infrastructure, governance, creativity, and power.
Contact: liveaiwire@gmail.com | Follow @LiveAIWire