2023-11-29 DeepfakesIDV DG Meeting Notes DRAFT

Owned by
Andrew Hughes
, created
Last updated: Dec 13, 2023

Meeting status metadata

Quorum

quorate

Notes-Status

Ready for review

Approved-Link

 

The meeting status metadata table is used for summary reports - copy the status macros from the table in these instructions:

Quorum: quorate not quorate

Notes-Status: drafting Ready for review approved

Approved-Link: Insert a link to the Meeting Notes page holding the approval decision for this notes page

 

Agenda

  • Call meeting to order

  • Roll call & determination of voting participant quorum

  • Agenda bashing

  • Approval of previous minutes (11/01, 11/08, 11/10, 11/15)

  • Announcements

  • Main discussion items (see table below)

    • Status updates from content area leaders

    • Update from Beruku about deepfakes testing

  • AOB

  • Adjourn meeting

 Attendees

Voting participants

Voter

Organization

Presence

Voter

Organization

Presence

Andrew Hughes

FaceTec, Inc.

Present

Chris LaBarbera

Verizon

Present

Iain Corby

SafetyTech Limited

Present

Jay Meier

FaceTec, Inc.

Present

Jim Pasquale

ChasmCrossing Solutions

Present

Jordan Burris

Socure

Present

Justin Byrd

N/A

Present

Maxine Most

Acuity Market Intelligence

Present

Mike Magrath

Easy Dynamics

Present

Peter Davis

Airside Mobile, Inc

Present

Sean Lanzner

iProov

Present

Non-Voting Participants

Participant

Organization

Presence

Participant

Organization

Presence

Alastair Treharne

Beruku Identity Limited

Present

James Monaghan

n/a

Present

Julian White

Beruku

Present

Guests

Participant

Organization

Presence

Participant

Organization

Presence

Kay Chopard

Kantara Initiative

Present

Sal D’Agostino

ID Machines

Present

Quorum determination

Meeting is quorate when 50% + 1 of voting participants attend

There are <<15>> voters as of <<2023-11-29>>

Approval of Prior Minutes

Motion to approve meeting minutes listed below:

Moved by: Maxine

Seconded by: Peter

Link to draft minutes and outcome

Discussion

Link to draft minutes and outcome

Discussion

  1. https://kantara.atlassian.net/wiki/spaces/DGDF/pages/279511041 Approved
  2. https://kantara.atlassian.net/wiki/spaces/DGDF/pages/283443213 Approved

None

  1. https://kantara.atlassian.net/wiki/spaces/DGDF/pages/285769729 Approved

 

  1. Approved

 

 Discussion topics

Time

Item

Presenter

Notes

Time

Item

Presenter

Notes

 

Kantara updates

Kay Chopard

  • This DG work is generating lots of interest and attention at industry events - confirmation that this is a hot topic and in demand

  • Reminder about the AGM next week - please register to hear about all the activities of Kantara and strategic plans

 

Content area updates

A.1 Glossaries and concepts

Jay

  • Jay has posted material that he has written in 2006 - needs some updating

 

A.2 IDPV

Andrew

  • No progress - plan to use the ENISA report on remote proofing

  • Jordan will try to get some citable material on current attacks that are beating existing controls

  • All to gather and post reports and materials into the “Source Material” wiki area

 

A.3 AI

James

  • Added some good resources - YouTube lectures - they are good to learn from

  • This one is very good:

  • The open source face matching models explain how they work in their documentation

  •  

 

C.1

 

  • Dawid created a write up - everyone to review and comment please

  • Companies in this field - you are welcome to add sections/pages/content on countermeasures

 

Update from Beruku about their work

Julian White

  • Beruku already does biometric testing

  • Have been extending into deepfake detection

  • They are creating deepfakes and using them to attack systems - to see how the systems respond

    • One question they are working on: “Can a deepfake based on a photograph be created to successfully attack systems?”

      • This is a complex attack to develop - but once the algorithm works scaling is easy

  • Note that using mobile device requires security based compromises

    • Remote onboarding over web/HTML5 is very easy - there is no hardware security

  • Are there companies that compile statistics about categories of attacks

    • Is this something Kantara could spearhead?

    • Define a reporting format?

    • What kinds of categories would we want to have?

    • Look at threat reports

      • increase in emulator use

      • increase in face swaps

      • motion based attacks

    • Is there a common ‘toll gate’ that all systems pass through - that can be used to create common metrics

    •  

 Open Action items

Action items may be created inline on any page. This block shows all open action items from all meeting notes.

 

 Decisions