A new algorithm that can create realistic looking videos of speech from audio files has been announced in a paper published by the University of Washington.

It is getting more difficult to separate reality from fantasy and the rise of artificial intelligence seems to increase this dilemma – Researchers at the University of Washington recently announced a new video-editing tool which they used to superimpose audio onto a video of former U.S. president Barack Obama, making it appear as though he’s saying whatever they want him to (with astonishingly realistic lip movement).

This means anyone can use this to put words in Obama’s mouth by converting audio sounds into mouth movements, then blend the movements into some old video footage.
One troubling issue that quickly comes to mind, is that this could increase the potency of fake news. You can see a photo-edited image from the 2017 G20 summit that went viral on social media recently:


The image only highlights the cheap tricks made possible by new editing technologies.

Many advances in human mimicking computer programs have emerged in the last few years. Google DeepMind and Lyrebird can already fake voice recordings. However, this by no means spells the doom for media evidence but calls for equally outstanding counter-measure technologies to up the quick-verification game.

Interestingly, scientists behind the video-editing project give the assurance that the computer system could also be used as a reverse verification tool. This means people can use it to determine whether a video is real or not. They hope the research will eventually be used in Hollywood special effects or to improve the quality of video calls. See the video-editing footage below.

Links: University of Washington

Trans Energize empowers her readership with conversations on emerging possibilities. We explore new thinking in business, lifestyle and society. Enter any questions you may have as comments below. Our media team is eager to respond and share thoughts with you . . .