Source Material
- Andrew Hughes
- mchaudoin@mindspring.com
- PeterD (Deactivated)
Please post links to reference materials below.
URL | Site | Description of content | Â | Â |
|---|---|---|---|---|
https://media.defense.gov/2023/Sep/12/2003298925/-1/-1/0/CSI-DEEPFAKE-THREATS.PDF | NSA/FBI/CISA | Joint CSI | Contextualizing Deepfake Threats to Organizations | Â | Â |
https://www.dhs.gov/sites/default/files/publications/increasing_threats_of_deepfake_identities_0.pdf  | DHS report on deepfakes…. | Not sure when this was published, but its pretty thorough, describing/explaining deepfakes |  |  |
HTML version:  [PDF is loaded in Attachments tool below] | Government of Canada: Canadian Security Intelligence Service | “On 24 May 2023, the Canadian Security Intelligence Service (CSIS) Academic Outreach and Stakeholder Engagement (AOSE) and Analysis and Exploitation of Information Sources (AXIS) programs co-hosted a workshop to examine the complex set of threats posed by deepfake disinformation technologies.“ |  |  |
The Verge - May 2022 | Report about Sensity’s experience testing liveness detection systems and how many customers and vendors weren’t super-concerned about it. [obviously times have changed] |  |  | |
https://www.etsi.org/deliver/etsi_gr/SAI/001_099/011/01.01.01_60/gr_SAI011v010101p.pdf | ETSI Securing Artificial Intelligence (SAI); Automated Manipulation of Multimedia Identity Representations | Broad coverage of video, audio, and text. Includes very basic countermeasure tactics. | Â | Â |
Psychology Today blog (Jan 7, 2024): New Psychological and Ethical Dangers of 'AI Identity Theft' - Creating AI replicas without people's permission poses psychological risk. | Marlynn Wei M.D., J.D. discusses the psychological aspects to deepfakes. | Â | Â | |
https://www.politico.com/news/2023/12/22/arizona-ai-election-deep-fakes-00133113 | Politico, December 22, 2023 | Arizona creates own deep-fake election hoaxes to prepare for 2024 | NEWS - Arizona creates own deep-fake election hoaxes to prepare for 2024 Â | Â |
 Â | National Acadamy of Sciences report on Face Matching. | Not quite comprehensive analysis of the current state of the market for face matching, including tech development history, core functionality and error definition, technical deficiencies, legal/regulatory overview and recommendations. Â | Â | Â |
Gartner Forecasts Doubt in Biometric Solutions Amid Rising AI-Gene… (smestreet.in) | SME Street | By 2026, attacks using AI-generated deepfakes on face biometrics will mean that 30% of enterprises will no longer consider such identity verification and authentication solutions to be reliable in isolation, according to Gartner, Inc. |  |  |
PR Newswire | BERKELEY, Calif., Feb. 21, 2024 More than 300 international experts in technology, artificial intelligence, digital ethics, and child safety, as well as entertainment professionals and academics issued an open letter urging governments, policymakers, and political leaders to combat the growing threat of harmful deepfake content, which often involves sexual imagery, fraud, or political disinformation. | Â | Â | |
Au10tix researchers discover relaunched fake ID generation site OnlyFake | Biometric Update | Biometric Update | OnlyFake . Org web site caused a stir among digital identity verification providers and businesses relying on biometrics to onboard customers when it was revealed as a tool for attempting fraud at scale. The website vanished from the internet following the outcry. Fraud prevention researchers at Au10tix discovered the site, reborn with a new URL. The synthetic ID generator added more disclaimers that demonstrate their awareness of how customers can commit fraud with the products they sell. | Â | Â |
Consumer Trust in Biometric Data Security Wanes - PaymentsJournal | Payments Journal | Consumers have become increasingly skeptical about tech companies’ ability to protect biometric data. A new survey from GetApp, which polled 1,000 consumers, revealed that the percentage of respondents who trust tech companies to keep biometric data safe dropped from 29% in 2022 to 5% in 2024. |  |  |
A small section dedicated to deep fakes, and more general guidance provided for various attack vectors. | Â | Â | ||
MIT Technology Review | Article about AI startup, Synthesia in the UK that can generate realistic Avatars with a participant's consent. Interesting discussion about the technology used and where it might be going. | Â | Â |
Â
Use this table to post links to vendor-supplied content (such as marketing material)
URL | Site | Description of content |
|---|---|---|
Microsoft’s VASA-1 can deepfake a person with one photo and one audio track | Ars Technica | April 16, 2024 Microsoft Research Asia unveiled VASA-1, an AI model that can create a synchronized animated video of a person talking or singing from a single photo and an existing audio track. |
Â