By Laura Kobylecky
Some people see them as a fun Internet novelty while others may see them as a sign that could lead to the total collapse of media trust. Ian Beacraft, CEO and Chief Futurist at Signal And Cipher (an agency that helps brands adapt to future tech) has a more optimistic perspective.
Synthetic media is a broad term for the production of media through artificial means. Beacraft introduces the concept with some popular Internet animations. The videos show celebrities and movie characters moving their mouths in sync to the words he is saying, creating the illusion that they, rather than Beacraft, are the ones speaking. These videos are rudimentary. The mouths have slightly blurred edges and a certain dissonance that flags them as fake. However, the result is close enough to convey the idea.
Beacraft then brings up animated versions of himself speaking in different languages, followed by a moment where he is placed in that iconic Matrix scene and he takes the obligatory “red” pill. He further illustrates the concept with an image of Jennifer Lawrence accepting an award, but her face has been digitally replaced by Steve Buscemi. These are some examples of that synthetic media which is manipulated or made entirely by algorithms.
He addresses the more nefarious possibilities of this technology. This sort of digital replacement could be used to artificially create “revenge porn” or to alter statements made by political figures. He balances this with the more professional applications of synthetic technology. Paul Walker appeared in Furious 7, after his death, using a digital recreation. In Rogue One, a particularly young Princess Leia was brought back for a cameo.
Beacraft explains three categories of deepfakes: voice cloning, puppetry and face-swapping. Voice cloning involves taking an “imprint” of someone’s voice and then using that in a variety of contexts, different from the original instances. A face-swap is the digital replacement of one face with another. Puppetry involves making someone appear to say words that they did not, by digitally altering the movement of the face to match the words. All of these are in some ways based on existing media and existing people.