Vitalik Buterin Says AI Could Become Ethereum’s New User Interface

Ethereum co-founder Vitalik Buterin said AI could become a new way for users to interact with blockchain networks as zero-knowledge technology enables more private and programmable data.

Speaking Oct. 6 at the OKX NOW global product and ecosystem conference in Singapore, Buterin described AI as both a major opportunity and a growing security risk for blockchain systems.

He said zero-knowledge proofs and other cryptographic tools can let users prove facts about themselves, assets or other objects without disclosing complete information. He expects the technology to become widely used within two years, including in ways users may not notice directly.

Buterin said privacy will become more important as AI systems process more personal data. AI agents can help users think and act, but they may also gain access to confidential information, including material covered by attorney-client privilege or nondisclosure agreements.

He also said AI is improving at attacking systems and finding software vulnerabilities. AI has helped identify weaknesses in Ethereum’s protocol, support complex cryptography and perform formal verification at both the protocol and application levels, he said.

That could make stronger security standards achievable, but those standards will also become necessary as AI becomes better at finding flaws.

About a month earlier, Buterin used a locally run AI agent to write a script that updated his ENS Hash without a traditional interface. He expects users to perform increasingly complex blockchain operations through AI agents rather than conventional apps or websites.

The shift could reduce risks linked to web interfaces, including compromised front ends. It could also create new threats, such as prompt injection, mistaken interpretations of transaction instructions and uncertainty about whether an AI system can be trusted to act as intended.

Buterin cited an attack on Safe, one of the largest crypto wallets, that caused about $1.4 billion in losses. He said the attack targeted the layer between users and the blockchain rather than smart contracts or on-chain code.

Over the next two years, Buterin expects Ethereum transactions to become less expensive and more private. Some applications already help users pay for AI inference and improve privacy for those payments.

He said bots could increasingly become users’ counterparties and application developers. Bots could publish offers on a blockchain, while a user’s bot interprets them and recommends which to accept.

The security challenge could eventually extend from software and protocols to the hardware that stores and connects digital systems to the physical world.