Deepfake monitoring system presented at SPIEF
As part of the session of the St. Petersburg International Economic Forum, ANO “Dialogue Regions” presented a system for monitoring deepfakes in real time, TASS reported on June 16, citing the press service of the organization. The development was called “Zephyr”.
Zephyr is a real-time transcription-based audiovisual monitoring information system that instantly detects deepfakes through algorithmic evaluation and analysis using artificial intelligence.
“Dialogue Regions” noted an increase in the amount of fake information in the media field. So, from January to May 2023, 1.6 thousand unique fakes and five million copies of them were identified, the organization specified. Growth compared to the same period last year amounted to 10% and 42%, respectively.
The appearance of the system was “a decision whose necessity is obvious,” Alexander Khinshtein, head of the Duma Committee on Information Policy, Information Technologies and Communications, responded to the Zephyr’s presentation. He stated that deepfakes are “not only someone’s glued face in defamatory videos,” but also methods of stealing accounts, funds from bank accounts, and also “one of the weapons in the information war.”
In turn, Anton Gorelkin, Khinshtein’s deputy on the committee he heads, called for “as widely as possible” to implement systems to search for fake information. The parliamentarian believes that they should be “integrated into domestic social networks and video hosting sites” so that the system checks the content when uploading and informs users about the presence of deepfakes.
