Ethereum's Infrastructure is Critical in Building Trust and Safety in AI
summary
Claims that Deep Prove can generate 1,100x faster zero-knowledge proofs and is the only commercially viable library that proves LLM inference, supporting frontier models (300–400B parameters) and GPU hyperparallelism. Demonstrations include verifiable inference over committed ONNX models, deployed with partners like Intel, Nvidia and AWS, plus live use in defense, hospital diagnostic systems and content-moderation workflows. Roadmap items: proofs of training, fairness and composable reasoning as ZKML moves beyond Web3 into regulated sectors.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:36Zero Knowledge Proofs & ZKML Overview
- 1:35Challenges in AI Trust, Safety, and Bias
- 2:26Importance of ZKPs for AI Verifiability
- 3:17Lrange Labs & Deep Prove Introduction
- 4:29How Deep Prove Works
- 5:56Use Cases: Defense, Healthcare, Web2, Web3
- 8:05Decentralization & Open Source Approach
- 9:07Lrange Research & Industry Collaboration
- 10:44Extending ZKML Beyond Crypto
- 11:29Societal and Sectoral Impact of ZKML
- 13:14Agentic Systems, Compliance, Content Moderation
- 15:02DeFi, Governance, and Healthcare Applications
- 16:40The Future: ZK as a Default Primitive
- 17:43Q&A: ZK Computational Improvements & Future Directions
- 20:33Conclusion
related talks
16mTruth in a Post Truth World
44mGPU-less, Trust-less, Limit-less: Reimagining the Confidential AI Cloud
40mEthereum at Hyperscale
15mBuilding AI Agents the Web3 Way
16mBuild an Onchain AI dapp with 0G's Inferencing
16mAI Without Limits: How ZeroG is Building the Future of Decentralized AI
21mKeynote: Ethereum and AI: Synergies and Potential Applications
37mEthereum Community and Values