+3.14 Digital Injection Prevention and Forged Media Detection

3.14 Digital Injection Prevention and Forged Media Detection

Many emerging attacks on both attended and unattended remote identity proofing processes pair digital injection attacks with increasingly effective and available generative AI tools. These AI tools are used to create or modify media that contain images or videos of applicants and evidence (i.e., deepfakes) to defeat automated document validation processes, biometric operations, and visual comparisons done by proofing agents. Injection attacks insert modified or forged media between the capture point (e.g., a device) and the element conducting the comparison or other operation (e.g., a server running the algorithms, a workstation used by a proofing agent).

All types of remote identity proofing are in some way vulnerable to these forms of attack, whether the attack is on the remote optical capture and inspection components, the automated biometric mechanisms, or the video systems used in remote attended processes.

A biometric comparison performed with a captured sample does not prevent these attacks. However, live document capture and presentation attack detection mechanisms do provide some protection from injection and forged media attacks by making the injection of viable forged media more challenging. Not only does the media need to be inserted into the communication channel between the applicant endpoint and the CSP comparison component, but the forged media would also need to sufficiently defeat any passive or active presentation attack detection mechanisms implemented by the CSP. However, even these mechanisms are not sufficient to address all possible cases of these kinds of attacks.

The following requirements apply to all remote identity proofing processes (i.e., unattended and attended) that make use of optical capture and recognition tools for evidence validation, remote biometric capture, and video sessions:

CSPs SHALL implement technical controls to increase confidence that digital media is being produced by a genuine sensor during the proofing process (e.g., detect the presence of a virtual camera, device emulator, or a jailbroken device).
CSPs SHALL analyze all digital media submitted during the identity proofing process for artifacts and indicators of potential modification, manipulation, tampering, or forgery. Automated image analysis algorithms SHALL be tested against available attack artifacts (i.e., forged and manipulated images and videos) and genuine media to provide a baseline of performance and to determine the expected rate of false positives and false negatives generated by the system. The kinds of available attack artifacts that were tested and the corresponding false negative rates SHALL be documented and made available to RPs upon request. Algorithmic analysis of media and automated decisioning SHOULD be augmented by manual reviews to address detection errors.
CSPs SHALL only use authenticated protected channels for the exchange of data during remote identity proofing processes.
CSPs SHOULD introduce a passive means of detecting forged or manipulated media for all capture scenarios.
CSPs SHOULD authenticate capture sensors or implement device attestation to increase the confidence in a device being used to transmit digital media as part of a remote identity proofing process.
CSPs SHOULD analyze digital media for signatures of generative AI algorithms and deepfake tools that are known to be used to create forged digital media.

The following additional requirements apply to remote attended collection scenarios:

CSPs SHALL train proofing agents and trusted referees to look for indications of manipulated media (e.g., high latency, synchronization issues, inconsistent skin tone and resolution).
CSPs SHALL introduce random “human-in-the-loop” cues into their capture processes to increase the possibility of forged or manipulated media being detected (e.g., by requesting user movements or requesting that the user move objects between the capture sensor and their face).

1. Visão geral

1.1 Referências

1.2 Requisitos identificados

1.3 Regulamentos relacionados

2. Requisitos identificados

Requisitos

3. Regulamentos relacionados

Regulamentos
  • NIST -

    EN: Publications of the National Institute of Standards and Technology (NIST), U.S. Department of Commerce, https://www.nist.gov/, from their electronic versions (OSCAL; SP 800-63-4 from its HTML version, https://pages.nist.gov/800-63-4/). Works of NIST employees are not subject to copyright protection in the United States (17 U.S.C. § 105); the OSCAL content is dedicated to the public domain worldwide (CC0 1.0); outside the United States NIST grants the right to reprint its technical series publications in all formats (NIST copyright statement). Republished courtesy of the National Institute of Standards and Technology; each document names its citation. Divided into sections and converted into the GRCA format, the wording unchanged. Only the versions published by NIST are authoritative.

    DE: Veröffentlichungen des National Institute of Standards and Technology (NIST), U.S. Department of Commerce, https://www.nist.gov/, aus ihren elektronischen Fassungen (OSCAL; SP 800-63-4 aus ihrer HTML-Fassung, https://pages.nist.gov/800-63-4/). Werke von NIST-Beschäftigten sind in den USA urheberrechtlich nicht geschützt (17 U.S.C. § 105); die OSCAL-Inhalte sind weltweit gemeinfrei gestellt (CC0 1.0); außerhalb der USA räumt das NIST das Recht ein, seine Technical-Series-Publikationen in allen Formaten neu zu veröffentlichen (Copyright-Hinweis des NIST). „Republished courtesy of the National Institute of Standards and Technology“; jedes Dokument nennt seine Zitierangabe. In Abschnitte aufgeteilt und in das GRCA-Format umgewandelt, der Wortlaut unverändert. Maßgeblich sind allein die vom NIST veröffentlichten Fassungen.