Sender identity verification project

What a Universal Identity certificate does to AI bot traffic

The same certificate architecture built to stop spoofed calls and emails also gives networks a way to tell a verified human or disclosed-bot sender from an anonymous automated one — before that traffic ever reaches you.

The Gap Today

Automated senders — AI-driven call bots, message bots, and content-generating accounts — currently look identical to legitimate traffic at the network level. There's no equivalent of a "return address" that a receiver, ISP, or platform can check before deciding whether to let a bot-originated message or an actual call coming through. A Universal Identity certificate closes that gap by attaching a verifiable, revocable identity claim to the senders data, and by requiring that claim to verifed by enduser whether the sender is legtimate or not.

How enforcement plays out at the Enduser edge

1

Sender presents a certificate

Every outbound call, message, or data request carries a certificate issued through the global CA hierarchy and an ISP sub-CA, including a disclosed attestation of whether the sender is legimate, a verified business system, or a bot that is not.

2

The Enduser gateway checks the certificate and possible revocation status

The ISP-issued home gateway (modem, cellular gateway, cable box) checks the certificate's trust tier and revocation status centrally, the same way it would for any sender — no extra hardware required in the home.

3

Undisclosed or unverifiable bot traffic is flagged or dropped

Automated traffic that fails to disclose itself as a bot, or that presents no valid certificate at all, is treated the same as spoofed traffic under the fail-closed model — flagged, deprioritized, or dropped depending on the trust tier the receiver has chosen.

4

Abusive sender certificates get revoked network-wide

A bot certificate that crosses the abuse threshold — spam volume, fraud complaints, scraping violations — gets added to the shared cross-ISP revoked-sender database, so it's blocked everywhere at once rather than ISP by ISP.

Before and after

Without a Universal Identity certificate

  • Bots and humans are indistinguishable at the network layer
  • No standard way to disclose automated origin
  • Bad actors re-register and reappear instantly after being blocked
  • Enforcement is reactive and platform-by-platform

With a Universal Identity certificate

  • Every sender carries a disclosed, verifiable trust certificate
  • Receivers can choose to filter or deprioritize undisclosed automation
  • Revocation is shared across ISPs, closing the re-registration loophole
  • Enforcement happens at the endusers edge, before traffic reaches the receiver

Extending the framework to AI bot traffic

This builds directly on the STIR-SHAKEN-style attestation and revocation model already proposed for calls and messages — extending trust-tier attestation to explicitly cover disclosed automated senders. Get involved or read the full technical framework.

Read the problem statement Support this work