Are we entering the new era of distorted reality, or are we already in it? With great advancements in deepfake technology, comes greater responsibility. The surge in deepfake technology has triggered a concern for the future of our privacy, democracy, and of course — our reality.
As we enter a new age in the digital world, we may start to question every piece of content we see and hear as deepfake technology is becoming more believable and accessible. While deepfakes can certainly pose opportunities for misuse, how will marketers adopt the technology responsibly to strengthen the relationship for brands to their consumers? On the other hand, what role does privacy play for consumers once the rise of deepfakes goes mainstream?
One thing we’re certain about, deepfakes are here for the long haul — so let’s get a better understanding on what they are.
What are Deepfakes
Deepfakes are synthetic media which can be in the form of audio, text, image, and video that is manipulated or generated commonly by deep-learning techniques such as machine learning and artificial intelligence. The term deepfakes was originally coined on the internet in 2017 by combining the terms “deep-learning” with “fakes”.
A form of deepfakes that haven’t garnered a lot of attention are AI-generated texts, which have been linked to being a highly deceptive form of deepfakes.
On the contrary, deepfake videos have made all sorts of headlines in the media. The following deepfake video released by MIT has sparked lots of conversation in the media as Richard Nixon famously doesn’t say the words “The men that went to the moon to explore in peace will stay on the moon to rest in peace”.
The video shown in this example demonstrates the high-level capability the technology has become to lip-sync generated audio from an original source while simultaneously manipulating the video to produce a highly convincing and deceptive output.
While in the beginning, AI-generated synthetic videos (using DNNs — Deep Neural Networks) were designed to deceive the human eye from authentic videos, advancements in technology are now taking measures to produce deepfakes undetectable to detection algorithms. The emergence of sophisticated deepfakes calls for detection technology to keep up the pace — digital-forensics expert on deepfakes from the University of California at Berkeley Hani Farid explains “The number of people working on the video-synthesis side, as opposed to the detector side, is 100 to 1”.
How Have Deepfakes Been Used
The Good
Deepfake technology can be a unique tool for marketers in the non-for-profit sectors to create public awareness and spread messages effectively. In the marketing campaign for voice petition Malaria Must Die, London startup Synthesia created a video ad of football star David Beckham speaking nine languages to say Malaria Must Die. The success of the campaign is a perfect example on how deepfakes can influence people around the world to take meaningful action. The petition allowed Global Fund to pass its goal and raise US$14.02 billion.
CEO & Co-Founder of Synthesia Victor Riparbelli explains how his company is “using AI through what we call visual synthesis — exactly the same thing as you would do in Hollywood studio”. He goes on to explain that they “used the video to build a 3D model and we can then re-animate that so you can change language, we’ll even rewrite the script of the film”.
The Bad & The Ugly
While David Beckham’s message on ending Malaria in countries most vulnerable to the disease was positive, there are potential negative implications with deepfakes that are important to discuss. Vox Media reported that pornography makes up for 96% of the deepfakes surfing the web today. These videos are almost always non- consensual and are a serious breach to people’s privacy. Celebrities have been the biggest victim of the technology to date due to the large access to footage from their films and TV shows. The advancements in the technology are making deepfakes easier to produce as there will be fewer amounts of original sourced footage required to produce fakes from. These fraudulent acts are a big part of the reason why the discourse around deepfakes has often been perceived negatively.
Marketing Implications With Deepfakes
The inevitable rise of deepfakes will expand the scope for marketers to use the benefits that the technology has to offer. Deepfake technology can be utilized in an omni- channel marketing approach from TV, podcast/radio, internet, email, and billboards in the future due to its potential for brands to personalize consumer experience, add an experiential factor for consumers, and save production costs in advertisement. While we might be years away from the normalization of synthetic media for our personal use, let’s take a deeper dive into how deepfakes will shape the future of marketing.
Advertising & Content Creating
The advertisement that put the Twitter world in a frenzy mid April, was ESPN’s Kenny Mayne State Farm ad which promoted the Netflix documentary series The Last Dance. ESPN teamed up with creative marketing agency Translation to produced this entertaining deepfake ad.
While we can tell we are watching a deepfake video — based on all the obvious little hints — it’s important to think about the psychological implications of it. For someone who has never watched any clips of the real Kenny Mayne prior to viewing this ad, their knowledge and experience with Kenny Mayne as a person would solely be based off this deepfake video. Even if you are fully aware that this videos was faked, the perceived sensory elements from the video such as sound, lip movement, tone, body movement, and facial expressions would be perceived the exact same way if this were a real video — with the exception that you know it’s fake. This can be said about any of the other deepfake videos we have seen. The manipulated reality deepfakes create in the consumer’s mind is a key reason why the technology will have a strong role in the future of marketing. Marketers of the future could use deepfakes in their marketing practices to create hyper-personalized experiences for consumers and alter consumers perception in the faked reality by creating deep, and emotional messages tailored to the brand’s strategic elements.
Other advantages of using deepfakes in the advertising industry would be the ability to correct footage errors or alter the script without having to re-shoot footage. Recycling old footage and creating synthetic video or audios can be useful tools for nostalgia marketing campaigns that brands and advertisers could run in the future. Lastly, the use of synthetic media could save time, and production costs which is an excellent way to reduce marketing budget spend.
Synthesia also developed synthetic media through deep learning to create two separate advertisements for distinct brands using one piece of footage. Their advertisement featuring Snoop Dogg for MenuLog was sourced off their advertisement from their parent company, Cratww’s JustEat. This led Synthesia to provide a cost- effective solution to their client by using artificial intelligence. The advantages of using deep learning and synthetic media could lead the future of content creation.
Bringing a New Level of Personalization to Experiential Marketing
Deepfakes can bring a whole new meaning to experiential marketing. While experiential marketing has commonly been known to be implemented as live marketing or event-marketing in the physical outdoor environment, deepfakes can bring experiential marketing online, directly to consumers.
The British start-up company SuperPersonal AI has been working on their app which would allow consumers to scan videos of their face, upload the videos to their platform and within minutes be able to produce deepfake output replacing the model with the consumer.
Another example showcasing how tech fashion has intriguing opportunities with the future of technology is through Microsoft SWAPP. The Fashion Innovation Agency (FIA), part of the London College of Fashion, collaborated with Microsoft to merge innovative technology for the fashion industry. Students Helen Wang MSc and Joanna Lanceley MSc founded Microsoft SWAPP which is the program designed to allow consumers to swap their faces into fashion ads, short films, and gifs that are tailored to a brand’s style & messaging. Powered by Microsoft Azure, the program uses machine learning and artificial intelligence to build realistic deepfakes for consumers to “see themselves” in the ads.
Lanceley states that “84% of millennials don’t like or trust traditional marketing” which might explain why they’re keen to host a disruptive platform which would allow brands to adopt experiential marketing as part of their visual commerce.
Interestingly, Tangent.ai is a software company that uses artificial intelligence by gathering information on consumers facial features and personalizes it based on the algorithms designed to provide the products that would look best on them. For example, the consumer can swap the model for to change their lipstick, hair colour, ethnicity, or race. As the study by Epsilon indicates that consumers are 80% more likely to make a purchase when brands offer personalized experiences, it is undeniable that brands must adopt personalized marketing strategies as much as they can.
Deepfakes provide a unique opportunity for personalization that will likely have a serious impact across the retail industry.
As the technology emerges, soon we might be scanning videos of ourselves on secure retail web-platforms and shopping for clothes, fashionable wear, and makeup. Next, we’d see the reflections of our faces in advertisements for products all over the webpage. Are ads of ourselves on shopping websites, retail marketing’s next big evolution? Retail brands need to be ready for deepfake trends entering the market and need to constantly look for ways new technology could play a role in their marketing strategy.
Digital Assistants & Influencer Marketing
Synthetic speech is becoming much easier to create as research suggests voice cloning of an individual can be reached via deep learning with only 5 seconds of speech, and 3.7 seconds with Baidu. The emergence of the voice cloning market is parallel to the rapid rise in usage of digital voice assistants (such as Google Home, Amazon Alexa, Siri). The demand in voice technology and voice cloning can have interesting marketing implications. As digital assistant devices can influence consumers daily decisions such as shopping, choosing a nearby restaurant, or delivering food, CMO’s might need to start thinking about marketing to digital assistants.
Google Assistant’s use of John Legend as their digital voice assistant is interesting to consider for the future of digital voice assistants and influencer marketing. Influencer marketing has proven to be an effective mediator for brands to connect with consumers on social media. By 2022, it is expected that brands spend up to $15 billion on influencer marketing according to Business Insider Intelligence data from Mediakix. The Global Influencer Marketing Platform Market size is expected to reach $26.4 billion by 2025. The growth in the influencer market suggests the growing impact influencers have on consumer purchasing decisions.
The rise in dependence for digital assistant apps and the use of audiofakes becomes even more interesting if we see people start to use influential people or those they actually know in their social circle, as their digital assistant voice. While we are now engaging in similar themes as the plot line of Spike Jonze’s 2013 film Her, how do we decide if voice clones of relatives, celebrities, or departed loved ones, are allowed to become a digital embodiment through voice technology on digital assistants?
Moreover, do brand’s marketing strategy to digital assistants become all the more important once an over-reliance on digital assistants occur?
Future of Marketing and Deepfakes
From the potential of audiofaked digital assistants to personalized retail websites with images and advertisements of yourself wearing all kinds of branded clothing, we are entering the world of deepfakes in which will be almost impossible to avoid. What remains to be seen is if brands become legally obligated to be transparent in their use of deepfakes. In that case, what would it look like? Perhaps, a universal synthetic media watermark to indicate a piece of synthetic media will become the norm?
It is imperative for detectable deepfake technology to maintain its pace with the development of deepfake technology. Marketers and the brands of the future have the opportunity to use deepfake technology to serve consumer needs and demands through safe and responsible use — but we must ensure that the technology is being used ethically, to truly be able to reap its benefits. Deepfake technology gives power, and as we know, with great power comes great responsibility.
Written By:
Sasan Navidi
Social Media Specialist |Home Trust Company
Past Editor: Future of Marketing Magazine
Graduate: Schulich Master of Marketing Program