Ming Guo, co-founder of LazAI, believes the convergence of AI and blockchain will reshape data privacy, model training, and AI development.
In a conversation with TheStreet Roundtable host Malak Albaw at TOKEN2049, Guo shared his vision for decentralized AI, highlighting the importance of small, efficient language models and how LazAI is tackling the challenges of centralized AI systems.
LazAI is a decentralized AI platform founded in 2023 by Ming Guo and a team of blockchain and AI researchers. The company focuses on building small, efficient language models powered by user-owned data and decentralized infrastructure.
“The current state of AI is overly centralized,” Guo said. “Big corporations dominate the space, and smaller entities have little to no opportunity to participate in training models or using data in a meaningful way. We want to change that.”
LazAI’s innovative approach is rooted in blockchain technology, ensuring that data used for AI training is not only transparent and secure but also controlled by the individuals who generate it. By decentralizing both data ownership and AI model training, LazAI aims to create an ecosystem where smaller entities can play a pivotal role in shaping AI development.
A new era of decentralized AI
At the heart of this transformation is LazAI’s use of Data Anchoring Tokens (DATs) and Individual-centric Decentralized Autonomous Organizations (iDAOs). These tools empower data providers and AI developers to retain control over their assets, ensuring that contributions are rewarded and the development of AI is more inclusive.
One of LazAI’s core beliefs is that the future of AI lies in smaller, more efficient models. “In the traditional model, we’re seeing enormous, centralized models that require significant resources,” Guo noted. “The decentralized model we’re developing allows smaller players to enter the game, enabling a broader range of participants to contribute to AI training and development.”
These smaller models can be trained by smaller entities, using decentralized infrastructure rather than centralized data centers. This allows LazAI to ensure data privacy and maintain control, giving the users — not the big corporations — ownership over their data.
Blockchain's role in AI development
Guo also emphasized the critical role blockchain plays in ensuring that data used in AI training is secure and verifiable. “Blockchain brings transparency to the process,” he explained. “Every piece of data used to train AI models can be traced, ensuring it’s used with permission and proper attribution.”