A system for securing a content processing pipeline includes a computing platform having a hardware processor and a memory storing a software code. The hardware processor executes the software code to insert a synthesized test image configured to activate one or more neurons of a malicious neural network into a content stream, provide the content stream as an input stream to a first processing node of the pipeline, and receive an output stream including a post-processed test image. The hardware processor further executes the software code to compare the post-processed test image in the output with an expected image corresponding to the synthesized test image. and to validate at least one portion of the pipeline as secure when the post-processed test image in the output matches the expected image.
BACKGROUND
Advances in machine learning have enabled the production of realistic but forged recreations of a person's image or voice, known as “deepfakes” due to the use of deep artificial neural networks for their creation. Deepfakes may be produced without the consent of the person whose image or voice is being used, and may make the person being represented appear to say or do something that they have in fact not said or done. As a result, deepfake manipulated digital content can be used maliciously to spread misinformation.
Due to the widespread popularity of digital content for the distribution of entertainment and news, ensuring data security of a processing pipeline for the digital content is vitally important to its creators, owners and distributors alike. However, as machine learning models continue to improve, detection of malware capable of producing deepfakes and introduced into one or more processing nodes of a content processing pipeline will continue to be challenging. As a result, in the absence of a robust and reliable solution for assessing content processing pipeline security, subtly manipulated or even substantially fake digital content may inadvertently be broadcast or otherwise distributed in violation of contractual agreement or regulatory restrictions, thereby subjecting the content owners and/or distributors to potential legal jeopardy. SUMMARY
There are provided systems and methods for securing a content processing pipeline, subs
1. A system configured to ensure data security of a content processing pipeline including a plurality of processing nodes, the system comprising: a computing platform including a hardware processor and a system memory; a software code stored in the system memory; the hardware processor being configured to execute the software code to produce a security assessment of at least a portion of the content processing pipeline by: inserting a synthesized test image into a content stream, the synthesized test image being configured to activate one or more neurons of a malicious neural network; providing the content stream including the synthesized test image as an input stream to a first processing node of the plurality of processing nodes of the content processing pipeline; receiving an output stream from one of the first processing node or a second processing node of the plurality of processing nodes of the content processing pipeline, the output stream including a post-processed test image; comparing the post-processed test image in the received output stream with an expected image corresponding to the synthesized test image; and validating at least one portion of the content processing pipeline as secure when the post-processed test image in the received output stream matches the expected image. ||
9. A method for use by a system including a computing platform having a hardware processor and a system memory storing a software code to produce a security assessment of at least one portion of a content processing pipeline including a plurality of processing nodes, the method comprising: inserting, by the software code executed by the hardware processor, a synthesized test image into a content stream as a test image, the synthesized test image being configured to activate one or more neurons of a malicious neural network; providing, by the software code executed by the hardware processor, the content stream including the synthesized test image as an input stream to a first processing node of the plurality of processing nodes of the content processing pipeline; receiving, by the software code executed by the hardware processor, an output stream from one of the first processing node or a second processing node of the plurality of processing nodes of the content processing pipeline, the output stream including a post-processed test image; comparing, by the software code executed by the hardware processor, the post-processed test image in the received output stream with an expected image corresponding to the synthesized test image; and validating, by the software code executed by the hardware processor, at least the one portion of the content processing pipeline as secure when the post-processed test image in the received output stream matches the expected image. ||
17. A computer-readable non-transitory medium having stored thereon instructions, which when executed by a hardware processor, instantiate a method comprising: inserting a synthesized test image into a content stream as a test image, the synthesized test image being configured to activate one or more neurons of a malicious neural network; providing the content stream including the synthesized test image as an input stream to a first processing node of a plurality of processing nodes of a content processing pipeline; receiving an output stream from one of the first processing node or a second processing node of the plurality of processing nodes of the content processing pipeline, the output stream including a post-processed test image; comparing the post-processed test image in the received output stream with an expected image corresponding to the synthesized test image; and validating at least one portion of the content processing pipeline as secure when the post-processed test image in the received output stream matches the expected image.