Where “Seeing Is No Longer Believing”, Is Deepfake A Destructive Machination Or Marketing Arsenal?
--
Implications Of Deepfake Technology on Brands
Barack Obama called Donald Trump a “complete dipshit”. Mark Zuckerberg bragged about having “total control of billions of people’s stolen data” to manipulate and exploit its users. Yet all this is not real. Deepfake technology is here to detach us from the truth.
Deepfake, using machine learning and deep learning, refers to any artificial intelligence-based human image synthesis technique of superimposing one face over another.
The algorithm from deepfake technology apps such as DeepArt, Wombo, Face Swap Live and AvengeThem allow users to swap faces onto movie characters, replace, animate faces or create AI-powered lip-sync videos. For example, users can easily transpose their own face onto Leonardo DiCaprio’s or “Game of Thrones” actor Kit Harington using the Chinese app Zao.
How does deepfake technology work?
Known as a Generative Adversarial Network (GAN), deepfake technology uses two separate neural networks to compete against each other: the Generator generates new pixel imagery mathematically similar to the existing images. The Discriminator is fed with another set of photos and tasked to identify synthetic photos. They compete and hone one another’s capabilities until eventually the Discriminator’s classification success rate falls to 50%, thereby producing sophisticated artificial images.
What Is A Shallowfake?
Shallowfakes, on the other hand, are crudely-made videos that are either presented out of context or doctored with simple editing tools. A shallowfake video that reached millions slowed down Nancy Pelosi’s speech and made the US Speaker of the House sound as though she was drunkenly slurring her words. Such simple manipulations could be used to shape public perceptions.
How To Spot A Deepfake?
Poor-quality Deepfakes such as poor lip synching, patchy skin tone, badly rendered hair strands, jewellery and teeth; or flickering lights and inconsistent illumination are easier to spot. Blinking irregularities which were often a telltale sign were quickly corrected within months to eliminate spotting.
Why We Should Be Concerned About Deepfake
Unsurprisingly, the first use case to which deepfake technology has been widely applied is pornography. The AI firm Deeptrace found 15,000 deepfake videos online in September 2019.
A staggering 96% were pornographic and 99% of those mapped faces were of female celebrities such as Gal Gadot, Taylor Swift, Scarlett Johansson and others on to porn stars in what is termed “revenge videos”.
Beyond porn, there are plenty of spoof, satire, extortion and mischief applied across politics, industries and individuals for calculated gains. Audio can be deepfaked too, to create “voice skins” or “voice clones” of public figures.
Are There Beneficial Uses Of Deepfake Technology?
There are legitimate and beneficial uses of the technology.
· Voice-cloning Deepfakes can recreate any voice for those with Motor Neurone Disease (ALS).
· Alzheimer’s’ patients can engage with videos depicting their younger selves and loved ones.
· Synthetic videos can bring historical figures to life in the classroom.
· Museums can create experiential interaction with dead famous painters. The Dalí museum in Florida has a Deepfake of the surrealist painter introducing his art and taking selfies with visitors.
· A film actor’s face can be placed on their stunt double’s body or used to resurrect dead actors.
Implications Of Deepfake Technology On Brands
1. Experiential Online Marketing
In the fashion industry, Tencent cited how Deepfakes can show outfits on models with different skin tones, heights and weights; providing a route for “very quick understanding” for customers viewing new collections from a brand.
2. Hyper-targeted Advertising and Dynamic Influencer Marketing
Gaming avatars are overlapping with real-life identities, while CGI models are mixing with real-life influencers. With Deepfake, influencers and celebrities can broaden their reach by fronting fashion ad campaigns and model clothes without turning up for a photo shoot; or have their voices reproduced in 16 different languages to achieve thousands of hyper-targeted ads.
A 2018 Zalando campaign which featured model Cara Delevingne, produced a range of Deepfake shots and voice fonts, gaining 180 million impressions across 12 countries on social media.
3. Hyper -personalised Retail Service
Retailers can replace a faceless bot with a deepfaked online assistant who is a customer’s exact demographic, speaking their language. Harvard Business Review revealed that when consumers see products as extensions of themselves, they are willing to buy more, pay higher and become brand advocates.
Will Deepfakes Wreak Havoc?
The problem with Deepfakes is not their method of creation, but rather, the ethical implications and specific harm causing harassment, shaming and intimidation.
The threat where “seeing is no longer believing” has serious systemic implications of a zero-trust society, especially when people cannot or no longer bother to distinguish truth from falsehood.
Despite the marketing positives, until our legislation catches up with the malicious use of Deepfakes, brands and individuals will have to be vigilant in their consumption, extraction of sources and production of video contents.
