Ethereum Founder Vitalik Buterin Builds ‘Paranoid Mode’ to Hide Your AI Secrets

An AI assistant helping manage investments could expose your holdings, trading plans, and financial questions to the company running its model. Ethereum co-founder Vitalik Buterin has built a way to make that trail harder to follow.

On October 16, Buterin unveiled a “paranoid mode” for his AI agent, designed to obscure who uses it, where requests originate, and how they are paid for. The code is public, but the combined setup remains experimental.

Wrapped zkAPI + Tor + extra cover traffic + local LLM prompt rewriting in an open-AI compatible daemon, and made a "paranoid mode" wrapper for my agent harness (pi):Each conversation gets assigned a separate Tor circuit, your requests are masked with random cover traffic, zkAPI… pic.twitter.com/2y236zA0cw

— vitalik.eth (@VitalikButerin) October 16, 2026

Why AI Agents Pose a Financial Privacy Risk

As AI agents gain access to wallets and financial tools, their conversations could reveal sensitive financial information. A trading assistant might disclose which tokens it monitors or details about a user’s portfolio.

Even anonymous users may be identified through payment records, internet addresses, or repeated details in prompts.

Buterin addresses those clues in four steps.

Tor gives each conversation a separate route to hide the user’s IP address. Fake requests disguise real traffic. A small AI model on the user’s computer rewrites prompts to reduce identifying details.

Finally, zkAPI uses cryptographic proofs to show that requests are paid for without identifying the payer.

The payment system already runs on Ethereum mainnet, accepting deposits in assets such as ETH and USDC. Buterin also confines his agent to a software sandbox to limit its access to his computer.

What This Means for Investors

Buterin says the system is “getting quite usable,” although each new conversation starts more slowly.

Complete anonymity remains elusive. Providers still see rewritten prompts, which could contain recognizable information.

He acknowledged another weakness: an AI provider cooperating with Tor’s first relay could study traffic timing and potentially identify real requests. He favors a more specialized privacy network, despite higher data costs.

definitely no point trying to de-link within a session – this daemon treats a session as a unit, and gains efficiency from *not* refreshing the Tor circuit or doing a zkAPI payment for each message within a sessioneach new session gets its own Tor circuit + zkAPI payment though

— vitalik.eth (@VitalikButerin) October 16, 2026

For investors, the longer-term question is whether private AI payments become cheap and fast enough for trading tools and other financial services.

Buterin says lower zkAPI transaction fees are coming. His prototype suggests a potential use for Ethereum infrastructure, but offers no evidence of widespread adoption or an immediate benefit to ETH’s price.