Sasha Ivanov (Units Network), Danny O’Brien (Filecoin Foundation), Mike Robinson (OORT), and Jayen Harrill (Covalent)
Our recent AI summit at DevCon explored pressing topics in decentralized AI, with a panel on DeAI & Censorship: Breaking Free From The Big Tech, delivering some of the most thought-provoking discussions of the event. Moderated by Jayen Harrill (Covalent), the panel included Danny O’Brien (Filecoin Foundation), Mike Robinson (OORT), and Sasha Ivanov (Units Network). Together, they tackled the complexities of ensuring ethical outputs in decentralized systems without relying on centralized oversight.
The panel opened with the nuanced challenge of integrating ethical safeguards into decentralized models while maintaining censorship resistance. As Danny O’Brien noted, “The place where censorship happens—or ethical constraints—is in the reinforcement learning layer. It’s about guiding the models without central control, creating a diversity of outcomes tailored to user needs.”
Mike Robinson highlighted how OORT addresses these concerns through community-driven mechanisms, saying, “We’re building infrastructure that is censorship-resistant, while empowering our community to uphold data integrity through staking and slashing mechanisms.” This innovative approach combines decentralization with active participation from users.
Sasha Ivanov brought a broader perspective to the table, emphasizing the delicate balance needed: “At the end of the day, there has to be some censorship—but it should be community-governed and decentralized. We must ensure public AI systems can function effectively while respecting privacy and local nuances.” Sasha’s insights highlighted the potential of federated learning as a blockchain-native solution to maintain privacy and data security.
From addressing regulatory challenges to exploring cryptographic innovations like ZK-Starks, the panelists left us with actionable insights and a clearer vision of decentralized AI’s future. The conversation reinforced the critical need for collaboration across the Web3 ecosystem to tackle these complex problems.
You can watch the full keynote on our YouTube here:
Comments