Ethereum co-founder Vitalik Buterin and Ethereum Foundation AI lead Davide Crapis have proposed using zero-knowledge proofs to anonymise user interactions with AI models while preserving safeguards against abuse.
In a joint blog post, the pair outlined a system designed to allow users to fund AI API usage privately while ensuring providers are paid and protected from spam or malicious activity.
“We need a system where a user can deposit funds once and make thousands of API calls anonymously, securely, and efficiently,”
They wrote, adding that providers must be guaranteed payment while users’ requests remain unlinkable to their identity.
The proposal addresses a growing tension between privacy and enforcement as large language models handle sensitive data, with current models relying either on identity-based access requiring emails or credit cards, or traceable on-chain per-request payments.
Under the suggested framework, users would deposit funds such as USDC into a smart contract and use zero-knowledge proofs and rate-limit nullifiers to make multiple API calls without revealing identity, while providers verify that cumulative usage remains within prepaid limits.
To deter abuse, the design includes a dual-staking mechanism in which double-spending attempts could forfeit a user’s deposit to claimants, while violations of terms of service would result in funds being burned and recorded on-chain for transparency.
“For example, a user might submit a prompt asking the model to generate instructions for building a weapon or to help them bypass security controls — requests that would violate many providers’ usage policies,”
Crapis and Buterin said, noting that while identities stay hidden, slashing events would remain publicly auditable.
At the time of reporting, Ethereum price was $1,960.45.