AI Passport

The AI Passport is a novel provenance system that provide the trust rails to the AI agent ecosystem. Much like tags and pedigree certificates for our pets, the AI Passport gives AI agents clear provenance certificates. When users interact with AI agents on apps or websites, they can verify the agent ownership, origins, and status—and make an informed decision about whether they can trust it.

The AI Passport features:

  • Privacy-preserving identity checks on those behind the agents — both AI Agent Service Providers and the agent owners
  • Unique AI Agent identifiers that can be used to trace origins and actions of each agent
  • Feedback system so consumers can rate or report agents
  • A TrustPilot-style AI agent review site
  • Control mechanism to manage agent permissions

The Passport Trust System

A highlevel explainer looks something like this:

A typical scenario looks like this:

  1. A user engages an AI service provider to create and manage an AI agent. This could be a chatbot that handles customer service on an ecommerce website
  1. Both AI service provider and the user (agent owner) verify their identity in privacy-preserving manner using concepts from personhood credentials and self-sovereign identity
  1. AI agent is assigned a Passport (i.e. a provenance certificate) and can now start performing its tasks on its stage, whether its a website, app, or phone call. Agent can be expressed in many modalities: programatic API calls, text-based chat, audiovisual avatars, or voice.
  1. When a consumer meets an AI agent, they can themselves verify the AI Passport on the agent directory—and also rate the agent or flag any issues they encounter.

Through this system, those behind a given AI agent can share who they are, while those who interact with the AI agent can verify who’s behind it. This is trust infrastructure like SSL certificates for websites.

Coming soon

The AI Passport is being built with care by Sebastian Nause-Blüml & Erasmus Hagen.

Get in touch! We want to hear from builders, thinkers, or anybody interested in unlocking the full potential of AI agents for the betterment of humanity.