2023-11-08 DeepfakesIDV DG Meeting Notes

Alastair Treharne

Beruku Identity Limited

Andrew Hughes

Ping

Present

Ashok Singal

Demystify Biometrics

Chris LaBarbera

Verizon

Present

Denny Prvu

RBC Royal Bank of Canada

Present

Iain Corby

SafetyTech Limited

Jay Meier

FaceTec, Inc.

Present

Jim Pasquale

ChasmCrossing Solutions

Jordan Burris

Socure

Julian White

Beruku

Justin Byrd

N/A

Present

Maxine Most

Acuity Market Intelligence

Present

Mike Magrath

Easy Dynamics

Present

Peter Davis

Airside Mobile, Inc

Dawid Jacobs

DAL Identity

Present

Heather Vescent

n/a

Present

Uttam Reddi

Aware Corp

Present

Brandon Gutierrez

Equifax

Present

James Monaghan

n/a

Present

Stephanie Schuckers

Clarkson University

Present

Dan Bachenheimer

Accenture

Present

Zarrak Khan

N/A

Present

Michael Engle

1Kosmos

Present (by AI Bot)

Announcement: New Survey

Heather V

• Created a survey related to the deepfakes topic

• Intended generally for Security and Identity industry people

• Expectation is that some of the outputs will come back into this group

• Based on Heather’s standard survey research method

• 1:1 interviews are possible to

• Please take the survey and circulate to your networks

Overview of work to date (wiki)

Andrew

If many newcomers to the call

Overview of AI Topic Area

James M

• James presents his initial thoughts on the AI topic area

• The challenge will be to cover the “AI” topics that are relevant to the IDPV topic - don’t cover all of AI - lots of that stuff exists and isn’t that useful to us

• Try to give people a place to hang their understanding on - e.g. the bullet about statistics → Data science → ML → AI - this is an evolutionary sequence that advances via new approaches, new technology, new names

• Generative AI exists because of the transformer

• James will post his document to the wiki soon

• Volunteers?

• Heather - has analysis tools/frameworks that might be helpful

• Dan B - there are additional core concepts that we should separate - uniqueness in a population, functional ID proofing, foundational ID Proofing, remote versus in-person IDPV

  • injection attacks (IT Security domain), Presentation attack detection (biometrics domain)

•  

Overview of Telephony topic area

Chris L

• Chris ran through slides about telephony and deepfakes

• The content is posted to the wiki already - please take a look

• Very interesting deck and topic - to detect attacks effectively is very infrastructure and cost intensive - so orgs must do a tradeoff analysis and smart sampling of suspected attacks otherwise costs are unsustainableB.3 Phone/Voice vector attacks and defences

Ideas for targeted invitations

Leads

Volunteers to self-select topic areas to join

All