Ethereum Foundation Launches Privacy Tool That Cuts the Link Between AI Prompts and Payers

The Ethereum Foundation has launched zkAPI, a system for paying for AI models and other metered online services without that usage being traced back to the payer.

The foundation’s dAI team announced the tool on Thursday, saying it was built with the Open Anonymity Project and is running on Ethereum mainnet. The project’s code repository describes the protocol as “experimental.”

Reaching an AI model through an API today typically means holding a key tied to an account and a payment method, which lets the provider connect every prompt to one person over time. “Prompts are personal. People ask AI models about their health, their finances, their doubts,” the foundation wrote.

Get Unchained’s crypto news in your inbox with the free Unchained Daily newsletter.

Users make a single deposit of crypto, such as ETH or USDC, into a vault contract on Ethereum. That balance then sits as a private note on the user’s device.

To spend, software on the user’s machine generates a zero-knowledge proof showing that some funded, unspent note covers the cost, without revealing which one. A zkAPI server checks the proof and issues a temporary API key with a dollar cap. Prompts then travel straight from the device to the AI provider with that key, and the private balance is charged for actual usage once the key expires.

“The server that handles money never sees content, and the provider that sees content never learns the billing identity behind a key,” the post said. Each spend also publishes a one-time identifier that exposes any attempt to reuse a balance. Users can withdraw leftover funds onchain even if every zkAPI server goes offline, the foundation said.

The client mimics the standard OpenAI and Ollama interfaces locally, so existing apps can plug in. Open Anonymity has also added zkAPI as a payment option in its OA chat app.

The foundation said zkAPI does not provide network anonymity. AI providers can still read what a request contains and see metadata such as a user’s IP address, and the post pointed users who want more protection to Tor. Repeated personal details or writing style in prompts can also let a provider re-link separate sessions.

zkAPI implements a design that Davide Crapis, who leads the foundation’s AI team, and co-founder Vitalik Buterin posted to the Ethereum Research forum in February. The launch comes nearly a year after the foundation set up a Privacy Cluster to coordinate privacy work across the network.

“It’s like a generalization of OA unlinkable inference that also abstracts payments away,” Ken Liu of the Open Anonymity Project wrote on X. Crapis said in a post that the team has “more things coming out to enhance user sovereignty and privacy in the age of AI.”

Related Listen: Zcash, Ethereum, Aztec, Canton and More: Which Chain Will Win the Privacy Race?

{"@context":"http:\/\/schema.org\/","@id":"https:\/\/unchainedcrypto.com\/ethereum-foundation-launches-privacy-tool-that-cuts-the-link-between-ai-prompts-and-payers\/#arve-youtube-ksqt1phwqii","@type":"VideoObject","embedURL":"https:\/\/www.youtube-nocookie.com\/embed\/KSQt1PHwqiI?feature=oembed&iv_load_policy=3&modestbranding=1&rel=0&autohide=1&playsinline=1&autoplay=0"}

The post Ethereum Foundation Launches Privacy Tool That Cuts the Link Between AI Prompts and Payers appeared first on Unchained.