On November 8th, 0x499 hosted a Twitter Space live event with the theme “Data Sovereignty and AI.” Six distinguished guests joined the online roundtable discussion: Scott Shi, CTO and Co-founder of Kite AI by ZettaBlock; Eric Cai, CEO of Entrova; Julian Peh, CEO & Co-Founder of KIP; Prashant Maurya, Co-founder & CEO of Spheron; Leanid, Head of Ecosystem at network3_ai; and Asa, APAC Head at io.net. Here are some highlights from the live event.

Moderator: Welcome everyone! Today we’re discussing data sovereignty and its impact on AI innovation.

Eric Cai: Hi, I’m Eric Cai from Entrova. I think data sovereignty is crucial, especially in AI, where systems depend on large datasets to learn and make accurate predictions. By respecting data sovereignty, we protect privacy, ensure transparency, and create AI solutions that enhance trust, rather than compromise user rights.

Moderator: Thanks, Eric. Can you tell us a bit more about Entrova?

Eric Cai: Of course! Entrova is building a decentralized, incentive-driven data platform designed to securely manage and optimize consumer data for AI training and application development. Our platform lets businesses and developers access high-quality, user-contributed data while enhancing privacy and security to create intelligent, personalized applications.

Moderator: Thank you, Eric. Next, we have Asa from io.net. Could you introduce yourself?

Asa: Hi! I’m Asa, Head of Asia-Pacific at io.net. We’re building the world’s largest decentralized GPU network, enabling anyone with surplus GPU power to contribute. This network provides access to GPU resources in a fast, permissionless way, which is essential as demand for GPU power grows. Currently, we’re generating over $20 million in revenue annually and expanding.

Moderator: That’s impressive! Could you share your thoughts on data sovereignty?

Asa: Sure! Data sovereignty is a fundamental tension in AI development. We need to aggregate data for models to learn and extract value, but centralizing this data can give big tech companies monopolistic power. Decentralization through blockchain can allow data to be aggregated while maintaining ownership and collective control, rather than simply handing it to corporations.

Moderator: Thank you, Asa. Next, we have Leonid from Network3.

Leonid: Hello, everyone. I’m Leonid, Head of Ecosystem at Network3 and a senior at UC Berkeley studying data science and finance. Network3 leverages IoT devices to train AI models locally, preserving data privacy. This decentralized approach allows AI agents to be trained on individual devices, ensuring data sovereignty since the data never leaves the user’s device.

Moderator: Thanks, Leonid. Prashant, would you like to add to this?

Prashant Maurya: Certainly. I feel that cross-border data restrictions can impact product quality and delay development in regions with strict regulations. However, as we explore synthetic data and federated learning, we’ll find ways to maintain compliance while driving AI innovation.

Moderator: Scott, as CTO of Kite AI by ZettaBlock, could you share your insights?

Scott Shi: Absolutely. Kite AI focuses on high-quality, domain-specific data, like medical imaging or multi-modal data, which are much larger and more complex than text data. Data sovereignty is critical for us because personal data is becoming increasingly regulated. Our approach focuses on providing infrastructure that tracks data ownership and attribution, ensuring contributors are compensated when their data supports AI models.

Moderator: Julian, from KIP Protocol, how do you see data sovereignty?

Julian Peh: Data sovereignty means being able to exert ownership rights over data, including the right to transfer or delete it. It’s also about guiding AI’s development to reflect the values of a specific culture or country. Sovereignty includes awareness of inherent data biases, so countries can ensure they aren’t adopting the biases of others.

Moderator: Great points. AI and big data have transformed industries, but data sovereignty poses unique challenges. How can AI help enforce data sovereignty?

Scott Shi: AI can help track data ownership, distribution, and lineage. However, most infrastructure isn’t designed for row-level tracking, so it’s difficult to monitor where data goes once it’s used in model training. That’s where we step in — our infrastructure tracks data lineage so that contributors can be fairly compensated based on their data’s role in model development.

Moderator: Eric, any final thoughts on AI and data sovereignty?

Eric Cai: Yes, data sovereignty laws like GDPR and CCPA set high standards for privacy and security, but navigating these regulations can be complex. AI will play a key role in detecting and managing compliance issues across sectors, ensuring that data is handled responsibly and securely.

Moderator: Thank you, everyone, for these insightful contributions. We’ll now open up the floor to audience questions.