In an era of AI and digitalization, we face a dual challenge in building trust while preserving privacy. For example, social platforms are flooded with unverifiable AI-generated content, and online services require users to excessively expose their personal data. Zero-Knowledge Succinct Proofs (ZK-SNARKs) offer a powerful cryptographic solution, enabling digital trust without compromising privacy. However, deploying ZK-SNARKs at scale requires overcoming fundamental obstacles: the significant computational overhead of proof generation and the security threat posed by quantum computing.
In this talk, I will present my research on improving the efficiency and security of ZK-SNARKs. I will first introduce folding schemes, a powerful primitive that significantly boosts the efficiency of proof systems. I will present LatticeFold, the first folding scheme that offers plausible post-quantum security, bridging the gap between high performance and long-term security. Building on this, I will briefly describe a framework that leverages folding schemes to bootstrap existing SNARK architectures for large-scale applications. I will conclude by outlining my vision for a future of verifiable and privacy-preserving digital infrastructure, discussing how the intersection of cryptography, quantum information, and AI can help us build a trustworthy digital world.
