AI, Knowledge Graphs, and Cross-Network Engagement: Enhancing Social Experiences in Web3 | Web3 Social Day Bangkok
Panel Guests:
- Yassine Landa, Founder & CEO, MBD.xyz
- Kameshwaran Elangovan, Core Contributor, Openledger
Moderator:
- Xiao Wu, Co-Founder, World3
Introductions
Xiao Wu (Moderator):
Hello, everyone. Welcome to our panel discussion on “AI, Knowledge Graphs, and Cross-Network Engagement.” We’re here to explore how these cutting-edge technologies are reshaping the Web3 social landscape. Let’s start with a brief introduction. Could each of you share who you are and what you’re building?
Kameshwaran (OpenLedger):
I’m Kameshwaran, one of the core contributors to OpenLedger. We’re building a blockchain for AI focused on data collection, pipelining, and attribution. We aim to enable more efficient data usage and monetization in AI systems.
Yassine (MBD):
Hi, I’m Yassine, the Founder and CEO of MBD. We build AI recommendation systems tailored for Web3 social media apps — essentially “AI as an API.” If you’re creating a Web3 Social app and want to compete with platforms like TikTok or Twitter, we provide advanced machine-learning tools for personalization and recommendations. Our mission is to build decentralized, algorithmic building blocks that empower users to control their feeds.
Xiao Wu (Moderator):
Thank you! I’m Xiao Wu, co-founder of World3. You can also call me Ling, my developer name. I’ve been coding since I was nine and started contributing to Ethereum about seven years ago. Let’s dive into our first question.
The intersection of AI and Web3 is a hot topic. Given your background, how is your work in this space creating unique cross-chain social experiences? How does it differ from traditional Web2 social platforms?
AI and Web3: Synergies and Innovations
Yassine (MBD):
While AI might seem like a hype topic now, it has deep roots. AI’s resurgence began with platforms like Facebook, which coined the term “data scientist” to process massive amounts of user data for recommendations. Web3 social faces similar challenges, such as spam and content overload, but with an open-data ethos that allows for decentralized and user-controlled AI.
Web3 offers new paradigms — open data, permissionless AI agents, and user-owned algorithms. These features not only enable better personalization but also ensure that AI systems are transparent and resistant to censorship.
Kameshwaran (OpenLedger):
Right now, in Web2, companies like Facebook and Google collect and monetize user data, offering entertainment or traffic in return. However, AI often operates unfairly — it consumes internet data without monetizing it for contributors. Our work focuses on data attribution, a key missing piece in the ecosystem.
For example, when training Large Language Models (LLMs), it’s difficult to trace which datasets influence the model’s outputs. By leveraging blockchain, we enable transparent data tracking, rewarding contributors whose data enhances AI models. This unlocks monetization opportunities for users while addressing the “black-box” nature of current AI systems.
Xiao Wu (Moderator):
Both of you bring up excellent points. It’s reminiscent of the early internet days when recommendation algorithms revolutionized platforms like Amazon. AI in Web3 has the potential to redefine personalization by merging these traditional techniques with blockchain’s transparency and openness.
The Role of Knowledge Graphs
Xiao Wu (Moderator):
Let’s move to the role of knowledge graphs. Could you explain how they organize decentralized user data for personalized interactions? How do they enable interoperability across Web3 networks?
Kameshwaran (OpenLedger):
A knowledge graph essentially tracks relationships between data points. For example, Story Protocol uses a knowledge graph to trace intellectual property (IP) usage. Imagine generating AI art of Spider-Man — it’s derived from an IP owned by Marvel. Knowledge graphs help track how such IP is used, enabling better monetization and compliance.
For our work, we focus on social data. Users generate valuable data daily, but individual datasets are often too small to train impactful AI models. By combining and synthesizing data — such as creating synthetic personas based on smaller datasets — we scale the impact and unlock monetization possibilities. Blockchain ensures this data’s transparency and attribution.
Yassine (MBD):
I see knowledge graphs as “behavioral graphs.” They’re essentially schemas that organize and connect different data types. In our case, we use them to build on-chain recommendations. For instance, when a user trades a meme coin, posts on Lens, or replies to a Farcaster thread, each interaction adds to a dense graph of nodes and edges.
Nodes represent entities — users, posts, NFTs — while edges capture relationships like “user follows,” “user mints,” or “user likes.” By layering advanced machine learning techniques on top of this graph, we predict user behavior and provide highly personalized recommendations. It’s similar to autocomplete but tailored to Web3.
Xiao Wu (Moderator):
Fascinating. Blockchain’s transparency and open-data nature make these graphs not only feasible but also incredibly powerful. Aggregating this data into structured, on-chain knowledge graphs offers immense opportunities for innovation.
Cross-Network Personalization and Privacy
Xiao Wu (Moderator):
How do you approach cross-platform personalization while ensuring privacy? How do knowledge graphs foster seamless Web3 experiences rather than isolating users?
Kameshwaran (OpenLedger):
The key is finding a balance between privacy and usability. Aggregating decentralized data and training it in controlled environments — like private small language models — allows for personalization without compromising user privacy. Tools like zero-knowledge proofs (ZKPs) can enhance this further, though scalability remains a challenge.
Monetizing open-source AI systems is another frontier. By tracking usage and attribution, we create opportunities for users to benefit financially from their contributions, even when using open data.
Yassine (MBD):
Web3’s pseudo-anonymity is a great feature. Users can separate identities with fresh wallets, enabling privacy by design. For cross-network personalization, cryptographic tools like ZKPs can allow users to selectively share data — such as YouTube history or past interactions — without exposing their entire profile.
By incorporating these selective disclosures into public knowledge graphs, we enhance personalization while respecting user autonomy. Blockchain’s transparent and immutable nature ensures that even these interactions are trustworthy and verifiable.
Xiao Wu (Moderator):
It’s incredible how blockchain redefines user privacy and data ownership while enabling innovative AI applications. Thank you both for these insights. This has been a rich discussion on how AI, knowledge graphs, and Web3 converge to enhance social experiences.
Balancing Privacy and Verification in Web3
Xiao Wu (Moderator):
Let me pose a slightly sensitive question. In Web3, anonymity is a core feature — you can create a fresh wallet and participate pseudonymously. But increasingly, especially in the Ethereum ecosystem, proving you’re part of the space often involves revealing certain things.
Take ENS names or transaction histories, for instance. Coders might link their GitHub profiles to Ethereum addresses to establish credibility or demonstrate they’re OGs. Yet this creates tension between staying anonymous and sharing details like monetization or trading history to build trust or reputation.
How do you think we can strike a balance between these competing needs of privacy and verification?
Yassine (MBD):
That’s a great question. I think it’s essential to distinguish between privacy and fairness. Privacy is contextual — sometimes people are fine sharing information. For example, if I’ve developed something or written an insightful post, I’m okay with making that public because I want credit for it.
Fairness, however, is where the lines blur. If I willingly share open data and someone else builds a machine learning model on top of it to make money without acknowledging or rewarding me, that’s unfair.
The solution? Implement revenue-sharing mechanisms for creators of public data. When we do this, we create incentives for people to contribute — even private data — because they know they’ll be compensated fairly. This fairness encourages a culture of participation.
Kameshwaran (OpenLedger):
Yeah, as long as we get paid.
Yassine (MBD):
Exactly! People are often willing to share if there’s fair compensation. For instance, if I watch a podcast and recommend it to my friends, I’m already sharing that information. But if there’s a way for creators to benefit from that sharing — say through a referral link or a revenue-share model — then I have more incentive to contribute.
Xiao Wu (Moderator):
Right, that’s an interesting point about incentivizing information sharing. It seems that Web3 has the potential to shift the incentive structure in more equitable ways.
Ownership and Monetization of Data
Yassine (MBD):
Monetization of data in Web3 is an interesting topic. Right now, data is locked behind closed walls — whether by corporations or centralized platforms. You only benefit from that data in a very limited way.
But Web3 opens up new opportunities. The idea is to empower people to own their data. You can monetize it in ways that weren’t possible before — through NFT minting, tokenization, and direct peer-to-peer transactions. This allows individuals to capture the value of their own data.
For instance, imagine that every time I contribute to an open-source project, I earn tokens that reflect my contributions. Those tokens could be used to buy services, vote on platform decisions, or even be traded. This is a big shift.
Kameshwaran (OpenLedger):
It’s a major step toward democratizing ownership. In the traditional world, all this data is siphoned off and controlled by centralized entities. But now, with Web3, we have the opportunity to regain control and extract value from our own information.
Yassine (MBD):
Exactly! It’s about making the flow of data more equitable, giving people the tools to share and monetize their contributions on their own terms. In a sense, Web3 transforms data from being just a commodity owned by platforms into a shared resource that benefits everyone.
The Role of AI in Web3 Social Experiences
Xiao Wu (Moderator):
I want to pivot slightly and talk about AI in Web3 social platforms. The integration of AI could potentially enhance user experiences by enabling more personalized interactions. What role do you see AI playing in this space?
Yassine (MBD):
AI can serve as a bridge between Web3’s decentralized nature and the need for personalized experiences. With AI, you can create recommendation systems that respect privacy and work across different networks. For instance, a user could opt-in to share certain aspects of their profile or activity, and AI would tailor content or connections based on that information.
The key is, unlike traditional systems where data is stored centrally, in Web3, this would be done in a more privacy-respecting way — using decentralized data storage or even zero-knowledge proofs to verify identities and transactions without revealing personal details.
Kameshwaran (OpenLedger):
AI will be key in bridging the personalization gap. Instead of relying on centralized databases, we’ll be able to use AI in a more decentralized way — whether through decentralized networks or edge computing — so that people’s privacy and control are still respected.
Yassine (MBD):
Right. It’s all about using AI to create more meaningful, personalized experiences without compromising on the principles of Web3, like privacy, ownership, and control over one’s own data.
Moving Beyond the “Web2” Social Model
Xiao Wu (Moderator):
How do you think Web3 social platforms can move beyond the typical Web2 social media model, which is highly centralized and profit-driven, into something that’s more community-oriented and decentralized?
Kameshwaran (OpenLedger):
The answer lies in creating incentives that benefit everyone in the ecosystem — not just the platform owner. With Web3, we can introduce new ways of rewarding users for their participation. Instead of being the product, users become co-owners or stakeholders in the platform.
For example, by introducing tokenized governance, users could vote on decisions and share in the platform’s profits. This creates a more balanced power dynamic.
Yassine (MBD):
Also, Web3 allows us to create decentralized identities (DIDs) that follow you across platforms. Your identity is not controlled by one centralized entity like Facebook or Google. Instead, you control your own identity, and you can choose how and where to share your data.
Web3 social platforms will allow people to take ownership of their digital selves, rather than having to conform to a one-size-fits-all model controlled by centralized platforms.
Xiao Wu (Moderator):
That’s a great point about decentralized identity. I think it’s crucial for Web3 to move beyond the “one-size-fits-all” mentality of Web2 and allow people to have more control and ownership over their digital lives.
