Skip to content
Deepak Gupta markFuture Tech

Future Tech/security

Your Agent Vets the Stranger's Agent by 2032

I spent years building the trust rails that let strangers transact at scale. The next layer is already forming: your agent vets the other person's agent before you ever say hello. Trust stops being a feeling you develop and becomes a credential you present.

// By 2032 · medium confidence · disruption 7/10

Prediction

// 2032

By 2032, verified digital identity and AI agents will mediate first contact between strangers, so your agent vets the other person's agent before you ever interact.

Confidence██████████medium
Disruption██████████7/10

What dies

  • → stranger danger
  • → the captcha

Who wins

  • → W3C Verifiable Credentials
  • → ISO mobile driver's license (mDL)
  • → Apple Wallet

filed: 2026-06-14 · guptadeepak.com

The hook

Right now you trust a stranger because an app shows you a 4.9 and a checkmark. But the app owns that score, and it knows everything about you to produce it. What happens when the score becomes yours, portable and cryptographic, and an agent presents it on your behalf? Trust stops being a vibe and a rating. It becomes a credential you hand over, and a negotiation between two pieces of software that happens before you say a word.

Thesis. The platform trust broker that killed stranger danger gets disintermediated by verifiable credentials and AI agents. Reputation becomes portable and cryptographic, and first contact between strangers becomes a vetting handshake between their agents.

The story

The setup: platforms own your trust today

We solved stranger trust by letting platforms vouch for strangers. The cost is that your reputation is trapped: your Uber rating does not travel to Airbnb, your eBay history does not help you on a new marketplace, and the platform can revoke any of it.

That trapped reputation is the soft spot. The same standards that fix login portability are coming for trust portability.

The hinge: trust becomes a credential, not a feeling

A verifiable credential is a signed, cryptographic claim you hold and present: this person is over 18, is a licensed driver, has completed 400 verified stays, has no fraud flags. You present the claim without handing over the underlying data, and the verifier checks the signature, not the platform.

Once trust is a portable credential, the human stops being the one who evaluates it. An agent does. Your agent reads the stranger's agent's credentials, applies your rules, and either greenlights contact or declines it before your attention is ever spent.

Current state: the rails are mostly built

Verifiable Credentials are a W3C standard. Mobile driver's licenses under ISO 18013-5 are live and in Apple and Google wallets. The EU digital identity wallet under eIDAS 2.0 is mandated and rolling out. Enterprise identity vendors already issue and verify credentials in production.

I built a CIAM platform I built that scaled to over a billion users, and the hardest part was never the cryptography. It was getting parties to accept each other's claims. That coordination problem is exactly what these standards and wallets are now solving in the open.

The trajectory: agent vets agent

By 2032 the default first interaction with a stranger, a date, a contractor, a co-host, a buyer, will be screened by an agent that checks credentials and reputation before surfacing the person to you. You will set a trust threshold the way you set a spam filter today.

Reputation will move with you. Leave a platform and your verified history comes along as credentials you control, not a score the platform deletes when you go.

The holdouts and the risk

The risk is a new monopoly. If a handful of issuers or wallet vendors become the only credentials anyone accepts, we have rebuilt the platform-as-trust-broker problem one layer down, just with better cryptography.

The optimistic case is genuine portability and selective disclosure: you prove what is needed, to whom you choose, without a single company owning the relationship. Which path we get depends on whether the standards stay open or the wallets become walled.

First signals (verify today)

The pieces are already shipping. W3C Verifiable Credentials reached Recommendation status and mobile driver's licenses (ISO 18013-5 mDL) are live in multiple US states and accepted by Apple Wallet and Google Wallet. The EU is rolling out its eIDAS 2.0 digital identity wallet. Major identity providers including Microsoft Entra Verified ID and Okta ship verifiable-credential issuance and verification. And AI agents that act on your behalf, negotiating and pre-screening, moved from demo to product across 2024 and 2025. [verify]

Key data points

  • W3C Verifiable Credentials Data Model reached Recommendation status, making it a formal web standard. [verify]
  • Mobile driver's licenses under ISO 18013-5 are live in multiple US states and supported in Apple Wallet and Google Wallet. [verify]
  • The EU eIDAS 2.0 regulation mandates a European Digital Identity Wallet for member states. [verify]
  • Microsoft Entra Verified ID and Okta both ship verifiable-credential issuance and verification in production. [verify]
  • Uber reported well over 100 million monthly active users and Airbnb over a billion cumulative guest arrivals, the scale that proves stranger-trust demand. [verify]
  • AI agents that pre-screen and negotiate on a user's behalf moved from demo to shipped product across 2024 and 2025. [verify]
  • Selective disclosure lets a holder prove a single attribute (over 18) without revealing the full credential.

Contrarian angle

The non-obvious part is who loses. The winner of the stranger-trust era was the platform in the middle, because it owned the rating and the identity check. Verifiable credentials and agents do not just help users, they disintermediate that broker: when reputation is portable and an agent verifies it directly, the platform stops being the thing you trust. The identity lens: you used to POSSESS a relationship you judged yourself, then you AUTHENTICATED into one a platform vouched for. The next step hands the credential back to you, but only if the wallets stay open instead of becoming the new gatekeeper.

The flip side

What this kills

The paired obituary in Tech Graveyard.

Read the obituary

FAQ

What does agent-mediated trust actually mean?

Before you interact with a stranger, an AI agent acting for you checks the stranger's verifiable credentials and reputation against your rules, and either approves contact or declines it. You judge the result, not the raw person.

How is this different from today's ratings and ID checks?

Today the platform owns the rating and the verification, and it does not travel. A verifiable credential is held and presented by you, signed cryptographically, and accepted by any verifier without routing through the original platform.

Why 2032 and not sooner?

The standards and wallets exist now but adoption and cross-acceptance take years. Driver's-license-in-wallet, eIDAS rollout, and agent products are all mid-deployment, so a default agent-vets-agent first contact is plausible by the early 2030s, not this year.

What is the biggest risk?

Trust monopolies. If only a few issuers or wallets are accepted everywhere, we recreate the platform-as-broker problem one layer down. Open standards and selective disclosure are the guardrails, and they are not guaranteed to win.

Does this finally bury stranger danger?

It buries what is left of it. The old instinct already gave way to platform ratings. Portable credentials and agents finish the job by making trust a machine-checkable claim, which is exactly why I tag the confidence medium rather than high.

More from guptadeepak.com

Want the technical deep-dive behind this prediction?

Read the companion article

Related predictions

More from the security desk.