BroadChain has learned that during an interview with CoinDesk at NEARCON2026 on March 5, Davide Crapis, AI Lead at the Ethereum Foundation, shared his perspective. He stated that as AI transforms sectors like finance and cybersecurity, Ethereum will not integrate with AI on the level of raw computational power. Instead, it will position itself as a coordination and verification layer in a world increasingly dominated by AI.
Crapis emphasized that if AI systems lack key attributes such as decentralization, autonomy, censorship resistance, and privacy—and if society becomes overly reliant on AI to manage everything—these crucial qualities could disappear.
Ethereum's AI strategy, he explained, is not about competing with giants like OpenAI or Google on model scale. The goal is to ensure that as AI becomes the primary interface for the internet, it does not inadvertently lead to a re-centralization of power.
This strategy is built around two core pillars:
First, decentralized AI coordination—providing the infrastructure for autonomous AI agents to establish identity, build trust, and facilitate payments. While AI computation itself remains off-chain, Ethereum would support agent discovery and verification through public registries, transparent historical records, payment routing, and cryptographic proofs. The related standard protocol, ERC-8004, is currently in development.
Second, embedding core principles—including privacy, openness, censorship resistance, and security—into the AI domain. This involves encouraging more AI processing to occur locally on users' devices, thereby reducing the amount of sensitive information sent to centralized servers.
Crapis also warned of future risks, noting that AI systems could potentially automate and scale up cyberattacks, challenging traditional authentication methods. In this context, he highlighted that cryptographic keys, due to their mathematically verifiable nature, will become increasingly vital.