Ethereum Foundation: Building the Ethereum network into an AI trust layer

In an interview with CoinDesk at NEARCON 2026, Davide Crapis, Head of AI at the Ethereum Foundation, stated that as AI reshapes fields such as finance and cybersecurity, Ethereum will not merge with AI at the raw computing power level, but rather act as a coordination and verification layer in an increasingly AI-dominated world. Crapis pointed out that if AI systems lack attributes such as decentralization, autonomy, censorship resistance, and privacy, and people rely on AI to handle everything, then these attributes will cease to exist. Ethereum's AI strategy is not to compete with OpenAI or Google on model scale, but to ensure that AI becoming the interface to the internet does not quietly lead to a recentralization of power. The strategy encompasses two main directions: First, decentralized AI coordination, providing the infrastructure for autonomous AI agents to identify, establish trust, and exchange payments. The AI's computational work remains off-chain, with Ethereum supporting agent discovery and verification through a public registry, transparent historical records, payment routing, and cryptographic proofs. The relevant standard protocol, ERC-8004, is under development. Second, it introduces core principles such as privacy, openness, censorship resistance, and security into the AI ​​field, encouraging more AI processing to be performed on users' local devices, reducing the amount of information transmitted to centralized servers. Crapis also stated that AI systems may automate and scale up cyberattacks in the future, challenging traditional authentication methods, while cryptographic keys will become increasingly important due to their mathematically verifiable nature.