AI didn’t break trust.
It revealed it was never solved.
We always relied on platforms to vouch for identity. That worked while platforms could scale human verification. It no longer does.
Two questions. No answer to either.
- 01
“Are you a real person?”
Generative AI can fabricate a face, a voice, a résumé, a review, or an entire online history. The signals we used to trust are now cheap to forge.
- 02
“Who does this AI work for?”
Agents act on our behalf across apps and organisations. Nothing today gives a verifiable answer to whose interests they actually serve.
Both questions are really one: trust has become the scarcest resource on the internet.
When a platform vouches for you, you’re trusting the platform.
Platform-mediated identity worked while a handful of companies could afford to verify humans at scale. It was always fragile: the vouching lived on someone else’s server, under someone else’s terms.
AI broke the economics. Verification that used to be expensive to fake is now trivial. And the fix can’t be another platform — a bigger gatekeeper is still a gatekeeper.
This isn’t a UX problem. It’s infrastructure.
- Healthcare Records that can’t travel — re-verified at every provider.
- Finance KYC completed five times for the same person.
- Democracy Verifiable participation that still preserves privacy.
- Commerce Agent-to-agent transactions with no way to verify a mandate.
- Communities No technical way to define your own membership rules.