Vitalik Buterin Proposes Human-Verified AI Wallets for Cryptocurrency

Buterin proposed AI-assisted wallets where algorithms suggest applications but users must confirm by making large transfers.

Vitalik Buterin expressed his opinion on how artificial intelligence (AI) can redefine the next generation of Web3 wallets.

He also proposed a model where people are always directly involved in authorizing high-value sales.

AI will shape new Crypto Wallets

The founder of Ethereum shared his thoughts on the respected social media site Farcaster, noting that it is “obvious” that the next iteration of wallets will heavily involve AI.

Despite this, Buterin added that he would not trust LLMs with multi-billion dollar transactions or controlling large sums of money. Instead, he provided a way in which AI systems assist users while leaving the final decision in human hands.

He described the ideal workflow for high-value scenarios that would involve an AI system proposing a plan, after which a local light client simulates the transaction. The person then reviews the intended action and the desired outcome before personally confirming it.

However, Buterin warned that this method should be used with care with a strong emphasis on security. He suggested that one way to achieve this is to remove links to decentralized applications from the transaction system. By removing the dApp user interface from the flow entirely, the system can reduce several attack vectors related to theft and privacy risks.

The 32-year-old has previously discussed how cryptocurrency and AI can evolve together. He envisions blockchains and technology working hand-in-hand, with crypto providing the trust, privacy, and economic infrastructure it needs to operate safely and impartially.

You may also like:

Proposed AI-Assisted Wallet Workflows

Other developers and community members responded to Buterin’s comments by explaining the potential uses of the idea.

Ethereum developer Andrey Petrov suggested two additional scenarios. In the first, the user starts the transaction as usual while the AI ​​analyzes the payment to be signed. The technology will then try to guess their intended action and explain it in plain language, allowing them to verify that the action accurately reflects what they intended to do.

In the second case, the user states his intended action directly or relies on the definition made in the first step. The AI ​​then tries to reconstruct the task independently, without reference to the actual value, to determine whether it reaches the same result. He explained that any disagreement between the two will indicate areas that need to be reviewed before the plan is finalized.

Another Farcaster user, identified as fkaany, described a framework in which AI programs complex crypto strategies such as multi-hop switching, yield optimization, and gas reduction.

This will include a local light client that simulates the result, which can allow people to review a clear summary and verify transactions manually, helping to reduce risks from blind signing, phishing links, and malicious dApp uploads.