Decentralized AI? Bridging Innovation with Ethics and Scalability

The Investing in AI Summit continued to push the boundaries of decentralized AI with a panel on DeAI & Censorship: Breaking Free From The Big Tech, a critical discussion on safeguarding innovation and integrity. Moderated by Jayen Harrill (Covalent), the panel brought together Danny O'Brien (Filecoin Foundation), Mike Robinson (OORT), and Sasha Ivanov (Units Network) to explore censorship, data integrity, and the balance between decentralization and ethical responsibility.

The discussion opened with a pointed question: How can decentralized AI ensure ethical outputs while resisting centralized control? Danny O'Brien from Filecoin challenged the assumption that "poisoned inputs" are the primary issue, emphasizing instead the importance of diversity in AI model outcomes: “One person's censorship is another person's desired output. Decentralized AI offers the opportunity for diversity rather than a centralized set of answers.” He highlighted Filecoin's role in maintaining data integrity through content addressability, ensuring data reliability in decentralized systems.

Mike Robinson from OORT emphasized their infrastructure’s focus on community-driven validation mechanisms. He shared, “Our proof-of-honesty consensus slashes validators for inaccurate computations, maintaining integrity while upholding decentralization.” OORT’s data collection and labeling platform further integrates community governance to align ethical standards with technical robustness.

Sasha Ivanov of Units Network underscored the inevitability of community-driven governance in censorship frameworks. He proposed federated learning as a blockchain-native solution for decentralized AI, explaining: “AI cannot function without some form of censorship, but decentralization allows this censorship to come from a community rather than a central authority.” He also elaborated on Units Network’s plans to integrate zero-knowledge (ZK) technologies and homomorphic encryption to bolster privacy and scalability.

The panel concluded with a nuanced view of decentralization’s role in addressing global compliance challenges, particularly in the face of regulatory pressures. As Danny O'Brien aptly summarized: “The more centralized systems lean into compliance at a global scale, the more decentralized systems can cater to nuanced, local needs without sacrificing freedom.”

The session offered an essential lens into the ethical, technical, and cultural challenges facing decentralized AI, leaving attendees with actionable insights on building resilient and responsible systems.

You can watch the full keynote on our YouTube here: