Skip to content

USE CASE · VIDEO CALLS AND DEEPFAKE RISK

Verify the person on the video call, regardless of how the video was made

Seeing a face no longer proves a person is present. Hearing a voice no longer proves a person is speaking. Deepfake detection asks whether media is synthetic; the more useful question is who is actually responsible for this interaction.

ACTOR TRUST RECEIPT

Illustrative
Actor
Human
Personhood
Verified
Identity
Verified privately
Presence
Continuous
Human control
High confidence
AI involvement
Low
Session continuity
High confidence
12:41 UTCact_91fk28...

In short

For video calls, Proof of Actor is designed to establish that the participant is the expected person, present and in control throughout the call, independent of whether the media stream could be synthetic, so remote meetings, interviews and approvals rest on evidence rather than appearance.

The question

Is the person on this call the expected participant, and are they present and in control right now?

Context

From detecting fakes to establishing actors

Deepfake detection tools judge whether audio or video appears manipulated. That is an arms race measured in model releases, and a clean result still does not tell you who is on the call.

Impersonation on video calls is used for payment fraud, fraudulent hiring and social engineering against help desks. The decision being made in each case, approving a payment, hiring a person, resetting credentials, depends on the participant's identity, presence and authority, not on pixels.

Proof of Actor is designed to attach actor evidence to the participant rather than the media: who they are, privately verified, whether they are present and continuous, and whether they are authorised for what is being asked.

Where video impersonation causes harm

  • Executive impersonation to authorise payments
  • Fraudulent candidates in remote interviews
  • Help-desk social engineering for credential resets
  • Remote identity checks and onboarding

What Proof of Actor is designed to establish

Trust layers this context needs

Each context asks for a different slice of the trust stack. These are the layers that matter here.

Identity

The participant is the expected person.

Presence

They are present on this call, now.

Actor continuity

The same person remains throughout.

Intent

Consequential requests are explicitly approved by them.

Policy

Typical policy for a payment approval call

Proof of Actor supplies the evidence. Your organisation sets and discloses the policy.

HUMAN (verified, present, explicit intent)
Proposed APIMeeting policy
// Proposed API - illustrative only
policy.require({
  actor: "human",
  identity: "verified_private",
  presence: "continuous",
  intent: "explicit_for_payment_requests"
});

FAQ

Frequently asked questions

How do you verify a person on a video call?

By establishing evidence about the participant rather than judging the video alone: that they are the expected person, verified privately, present and continuous during the call, and explicitly approving any consequential request.

Is Proof of Actor a deepfake detection tool?

No. Deepfake detection may become one signal among many. Proof of Actor asks a different question: regardless of how the media was produced, who is actually responsible for this interaction?

Can this prevent deepfake meeting fraud?

No system should promise prevention. Proof of Actor is designed to require evidence of identity, presence and intent before consequential decisions made on calls, which removes the reliance on appearance that such fraud exploits.

What interaction do you struggle to trust?

Design partners help define the evidence that would change their decisions. Tell us about yours.