Samsung Electronics has developed a new artificial intelligence technology to turn still images into a video to show various facial expressions and movements of a person.

According to industry sources and foreign media reports on Friday, Samsung’s AI lab in Russia published a paper to show the ability to turn artworks and celebrity photographs into ‘talking head videos’.

Unlike conventional "deep fake" videos, the technology jointly developed by Russia’s Skolkovo Institute of Science and Technology does not require a 3D modeling process and it can create a convincing fake video with just one only image.

The technology was unveiled to the public with its abstract published in the arxiv.org website for research papers in mathematics and physics. Experts say the technology can be widely applied to applications such as video calls, online games, and science fiction movies. It is also expected to be used to upgrade avatar images in Samsung’s smartphone Galaxy S10.

A researcher in the development team of the Samsung AI Center in Moscow uploaded a fake interview video on YouTube using the moving facial features of Marilyn Monroe and Mona Lisa.

Concerns are also growing about the potential uses of advanced AI technology in the creation of fake videos that can influence the public. Last year U.S. lawmakers began to propose a bill to regulate applications of such technology, warning faked videos could be damaging to national security. Facebook recently developed a machine learning model to identify fake videos.

Samsung Electronics opened the AI research center in Moscow last May to capitalize on Russia’s expertise in fundamental sciences. The Russian facility tapped experts in the machine learning field, including Professor Dmitri Vetrov of the Higher School of Economics and Professor Viktor Lempitsky of the Skolkovo Institute of Science.

By Hwang Soon-min and Minu Kim

[ⓒ Pulse by Maeil Business Newspaper & mk.co.kr, All rights reserved]