IOSG OFR13th Panel | The Art of On-chain Intelligence

Brought by:

🎙️ Michael Tung, Co-founder of Brevis
🎙️ Hilmar, Founder of Gelato
🎙️Ismael,  Co-founder & CEO of Lagrange
🎙️ Dylan Kugler, Ecosystem Lead of Starknet

In today’s rapidly evolving fields of machine learning (ML) and artificial intelligence (AI), ensuring the integrity, privacy, and trustworthiness of models has become a core challenge. Our discussion delved into how blockchain technology can play a transformative role in addressing these issues. This panel explored the key challenges in ensuring the reliability of machine learning models, particularly how to protect their integrity and credibility.

As privacy concerns become increasingly prominent, several cryptographic techniques—such as zero-knowledge proofs (ZKP), secure multi-party computation (MPC), fully homomorphic encryption (FHE), differential privacy, and trusted execution environments (TEE)—are being leveraged to enhance the privacy and verifiability of ML systems. Each method has its unique advantages and trade-offs, which we examined in detail.

One of the most promising areas is how blockchain technology can provide innovative solutions to these challenges. This is evident not only in enhancing transparency for training and inference processes but also in using zero-knowledge proofs to verify the provenance and integrity of ML models without disclosing sensitive information. We discussed how these technologies can address critical issues of bias, fairness, and trust in AI systems, as well as how they can assist in validating ML results in practical applications.