This was reported by Fox Business, according to Ukrinform.

The contract description states that the company must ensure “quick and accurate detection of deep frauds against the Russian and Chinese information warfare.”

“We are contracted to develop machine learning algorithms that can detect artificially created or modified faces or voices in all major languages, regardless of race, age and gender, and turn this AI into a platform that can be widely integrated into the Department of Defense. USA,” said Rijul, CEO of DeepMedia. Gupta.

According to him, generative AI and major language models help DeepMedia systems detect deep frauds.

“Our detection algorithms were trained on millions of samples, both real and fake, in 50 different languages, and we were able to tell with 99.5% accuracy whether something was real or was manipulated in some way by AI,” Gupta said.

DeepMedia COO Emma Brown noted that Russia’s war against Ukraine is a prime example of the importance of multilingual deep fake detection. In this context, he talked about the deepfake that surfaced at the start of a full-scale war with President Volodymyr Zelensky last year. “The only way Deepfake was detected was if linguists knew the accent was wrong. Thanks to our generative technology, we’re able to automate such moments in our system,” he said.

Deepfake is a method of synthesizing the image of a person based on artificial intelligence. It is used to combine some images and videos and overlay other images or videos.

As reported, on March 16, 2022, the Russians issued a “false statement by President Volodymyr Zelenskyi” that allegedly urged the Ukrainians to stop fighting and lay down their weapons. It was actively distributed by pro-Russian Telegram channels and separate sources, which were also attacked by hackers.