[DG-DeepfakesIDV] Re: How to detect AI deepfakes - The Washington Post

And to add fuel to the FUD fire, we see things like this: OpenAI says its voice cloning tool is too effective for public release. In

a blog post, OpenAI says results of testing show that its Voice Engine is so good at deepfake voice cloning and synthetic audio that it will certainly be misused on wide release, prompting the company to hold back on setting the product loose until it establishes stronger rules and guidelines for deployment.

Just in case there are politicians who do not fear the AI Deepfake Apocalypse - Altman gives them a (deepfake) bogeyman to worry about. I'd love to hear Chris LaBarbera's take on whether audio signal processing detection techniques can spot whatever the OpenAI innovation is. andrew. ———————— *Andrew Hughes *CISM m +1 250.888.9474 AndrewHughes3000@gmail.com On Fri, Apr 5, 2024 at 2:46 PM Terry Brenner wrote:

I picked up the pen to write the first draft on Prevention and Countermeasures against Deepfakes for the content pages. I will add non-liveness as a discussion point.

On Fri, Apr 5, 2024 at 5:13 PM wrote:

...

Apologies if I’m rambling on this one. ☹

What’s gray isn’t a part of the process and needs to be defined and built(I’ll explain why it’s critical), what’s in yellow is a data source we could have access too.

Michael F. Roche

FintechnologyGroup.com

Michael.f@fintechnologygroup.com

Cell: (216)346-9066

***Proprietary and Confidential

*From:* michael.f@fintechnologygroup.com < michael.f@fintechnologygroup.com> *Sent:* Friday, April 5, 2024 5:01 PM *To:* 'maxine most' ; 'Jay Meier' < jay@facetec.com> *Cc:* 'Olsen, Chris D.' ; 'Andrew Hughes' < andrewhughes3000@gmail.com>; 'Prvu, Denny (He/Him/His)' < denny.prvu@rbc.com>; dg-deepfakesidv@kantarainitiative.org *Subject:* RE: [DG-DeepfakesIDV] Re: How to detect AI deepfakes - The Washington Post

Just to be clear. non-Liveness/Biometric solution would be like KBA, 2FA/MFA, Security Tokens, SMS/Email verification, certificate/public keys, and captcha.

If so, I just saw a great demo on a solution that handles something similar but it’s after-the event.

Had me thinking…before the deepfake is made you do a facial biometric type captcha

- then when it’s done you go and make your deepfake. - Make the deepfake - Move to mobile authentication - authenticate on the phone with SMS > and the back of your drivers license. They have hooks into all the DMVs, and wireless carriers.

Only way to beat it would be ..I think

- Steal a real license id. - Then set up a cell phone in the name of who you’re trying to deepfake - Physically be at the address of the person you’re trying to deepfake

Michael F. Roche

FintechnologyGroup.com

Michael.f@fintechnologygroup.com

Cell: (216)346-9066

***Proprietary and Confidential

*From:* maxine most *Sent:* Friday, April 5, 2024 3:59 PM *To:* Jay Meier *Cc:* Olsen, Chris D. ; Andrew Hughes < andrewhughes3000@gmail.com>; Prvu, Denny (He/Him/His) ; dg-deepfakesidv@kantarainitiative.org *Subject:* [DG-DeepfakesIDV] Re: How to detect AI deepfakes - The Washington Post

I I'm not opposed to expanding to include this, but do we have someone who's willing to do the research and provide the content?

C. Maxine Most

Principal, Acuity Market Intelligence

www.acuitymi.com

+1 720 530 5836

On Apr 5, 2024, at 12:57 PM, Jay Meier wrote:

 I have a question about this content….

I’ve asked numerous time, if we should consider non-Liveness/Biometric solutions to the AI-DeepFake problem. This article is a good example of alternative solutions that we “could” include/address in our work.

Thoughts??

J

Jay Meier Senior Vice President - North American Operations jay@facetec.com | 612-978-3687

www.FaceTec.com | www.Liveness.com | www.SpoofBounty.com

Confidentiality Notice: This message is intended only for the use of the Addressee and may contain information that is privileged and confidential. If you are not the intended recipient, you are hereby notified that any dissemination of this communication is strictly prohibited. Please erase all copies of the message and its attachments.

On Apr 5, 2024, at 1:53 PM, Olsen, Chris D. wrote:

Thanks for sharing, Maxine!

The author (Gerrit De Vynck) is not conveying much more than the daily news hype in my opinion. There are a few nuggets in there but the article focuses on how difficult the issue is to solve primarily around image originality. What does this team wish to help solve? One of my concerns is how deep fakes will affect identity proofing and vetting.

Best,

Chris

*Chris Olsen | CISSP*

425-219-5987

Sr. Security Specialist, Identity & Access Management

Global Information Security | The Walt Disney Company

*From: *maxine most *Date: *Friday, April 5, 2024 at 11:29 AM *To: *Andrew Hughes , Jay Meier < jay@facetec.com>, Prvu, Denny (He/Him/His) , dg-deepfakesidv@kantarainitiative.org < dg-deepfakesidv@kantarainitiative.org> *Subject: *[DG-DeepfakesIDV] Re: How to detect AI deepfakes - The Washington Post

*This Message is From an External Sender* Caution: Do not click links or open attachments unless you recognize the sender and know the content is safe.

Free gift

https://wapo.st/49q5tHb https://urldefense.com/v3/__https:/wapo.st/49q5tHb__;!!Om87Lau1Cg!N84Lar2jxB...

*C. Maxine Most*

Principal, *Acuity Market Intelligence*

+1 720 530 5836

640 W. Linden St, Louisville, CO 80027, USA

www.acuitymi.com https://urldefense.com/v3/__http:/www.acuitymi.com__;!!Om87Lau1Cg!N84Lar2jxB...

https://urldefense.com/v3/__https:/www.linkedin.com/in/maxinemost/__;!!Om87L... https://urldefense.com/v3/__https:/twitter.com/cmaxmost__;!!Om87Lau1Cg!N84La...

*From: *Andrew Hughes *Date: *Friday, April 5, 2024 at 11:43 AM *To: *maxine most , Jay Meier , Prvu, Denny (He/Him/His) *Subject: *Fwd: [DG-DeepfakesIDV] How to detect AI deepfakes - The Washington Post

Paywall for me...

Are any names named? Should they know about our group?

————————

*Andrew Hughes *CISM m +1 250.888.9474 AndrewHughes3000@gmail.com

---------- Forwarded message --------- From: *maxine most* Date: Fri, Apr 5, 2024 at 7:55 AM Subject: [DG-DeepfakesIDV] How to detect AI deepfakes - The Washington Post To: dg-deepfakesidv@kantarainitiative.org < dg-deepfakesidv@kantarainitiative.org>

https://www.washingtonpost.com/technology/2024/04/05/ai-deepfakes-detection/ https://urldefense.com/v3/__https:/www.washingtonpost.com/technology/2024/04...

C. Maxine Most Principal, Acuity Market Intelligence www.acuitymi.com https://urldefense.com/v3/__http:/www.acuitymi.com__;!!Om87Lau1Cg!N84Lar2jxB... +1 720 530 5836 _______________________________________________ A Community Group mailing list of KantaraInitiative.org http://kantarainitiative.org/ DG-DeepfakesIDV mailing list -- dg-deepfakesidv@kantarainitiative.org To unsubscribe send an email to staff@kantarainitiative.org List archives -- https://mailman.kantarainitiative.org/hyperkitty/list/dg-deepfakesidv@kantar... https://urldefense.com/v3/__https:/mailman.kantarainitiative.org/hyperkitty/... ______ Group wiki -- https://kantara.atlassian.net/wiki/spaces/DFDG/overview https://urldefense.com/v3/__https:/kantara.atlassian.net/wiki/spaces/DFDG/ov...

_______________________________________________ A Community Group mailing list of KantaraInitiative.org http://kantarainitiative.org/ DG-DeepfakesIDV mailing list -- dg-deepfakesidv@kantarainitiative.org To unsubscribe send an email to staff@kantarainitiative.org List archives -- https://mailman.kantarainitiative.org/hyperkitty/list/dg-deepfakesidv@kantar... ______ Group wiki -- https://kantara.atlassian.net/wiki/spaces/DFDG/overview

_______________________________________________ A Community Group mailing list of KantaraInitiative.org DG-DeepfakesIDV mailing list -- dg-deepfakesidv@kantarainitiative.org To unsubscribe send an email to staff@kantarainitiative.org List archives -- https://mailman.kantarainitiative.org/hyperkitty/list/dg-deepfakesidv@kantar... ______ Group wiki -- https://kantara.atlassian.net/wiki/spaces/DFDG/overview