The Department of Defense is utilizing artificial intelligence (AI) to assist service members in identifying deepfakes that may jeopardize national security.
Modern AI technologies have made it incredibly easy to create convincing deepfakes, which aim to mimic the voice and appearance of a genuine person in order to deceive an unsuspecting viewer.
In the realm of national security, deepfakes could be employed to trick military or intelligence personnel into sharing sensitive information with an adversary who pretends to be a trusted colleague.
Recently, the Pentagon granted a contract to DeepMedia, a startup based in Silicon Valley, to utilize its AI-powered deepfake detection technologies.
The contract’s objective is to swiftly and accurately identify deepfakes to combat information warfare conducted by Russia and China, reported by FoxBusiness.
Machine learning algorithms to detect deepfakes
Rijul Gupta, the CEO, and co-founder of DeepMedia, stated that their company has been hired to create algorithms using machine learning.
These algorithms will have the ability to identify artificially generated or altered faces and voices in various languages across different races, ages, and genders.
What is #Deepfake technology? Well, Deepfake technology is an AI system that can create convincing but entirely fictional photos and videos from scratch. So, how do we know what's real and what's not? We'll find out more in this month's #CyberResearch#CyberSecurity #InfoSec pic.twitter.com/AAqoP6ZmKt
— CyberSec 17C (@CyberSec_17C) July 10, 2023
The ultimate goal is to develop an AI platform that can be seamlessly integrated into the Department of Defense (DOD) as a whole.
Gupta also highlighted the potential dangers faced by individual members of the Department of Defense, regardless of their rank or position.
He explained that anyone, whether they hold a high-ranking position or are at a lower level, could have their face or voice manipulated through deepfakes.
This manipulation could then be exploited by our international adversaries to obtain highly classified information.
DeepMedia’s deepfake detection systems
Gupta shared that their AI technology has the capability to analyze audio or video recordings and automatically extract the faces and voices present in that content. These extracted elements are then subjected to their detection algorithms.
The detection algorithms have been trained on a vast dataset consisting of millions of samples, encompassing both genuine and fake content in 50 different languages.
Moreover, these algorithms have achieved an impressive accuracy rate of 99.5% in distinguishing between authentic recordings and those that have been manipulated using AI techniques, Gupta claimed.
Emma Brown, the co-founder and COO of DeepMedia emphasized the significance of multilingual deepfake detection, citing Russia’s conflict with Ukraine as a pertinent example.
She mentioned a specific incident where a deepfake video featuring President Zelensky emerged during the early stages of the war in Ukraine.
In the manipulated video, Zelensky was portrayed as surrendering, but it was the linguists who detected the inaccuracy in the accent.
Ukraine 🇺🇦 : Facebook/Meta removed today a #deepfake video showing President Zelensky falsely encouraging Ukrainians to surrender.
➡️Maybe the more noticeable deepfake so far.
Apparently, Russian propagandists seek to interfere with the effective COM of the Ukrainian president. https://t.co/oaBPJMZc0z pic.twitter.com/l0pbU0VARn— Cyber, etc… (@cyber_etc) March 16, 2022
See all the latest news from Greece and the world at Greekreporter.com. Contact our newsroom to report an update or send your story, photos and videos. Follow GR on Google News and subscribe here to our daily email!
