What marketers need to be aware of about video fakes?
If you've heard the term "deepfake," you might immediately think of the fraud, ethically questionable or downright disturbing videos which have swept across the internet in the last several years. But as deepfake videos - and the AI technology that powers their creation - continue to become more advanced, it is essential that marketers and creators across all industries understand how they work and explore ways to use them to get and stay ahead of the changing world.
If you're not familiar with the term, deepfakes are synthetic media that has been altered digitally to substitute the appearance of a person convincingly, resulting in the ability to create videos that look and feel real however don't. Therefore, it's not surprising that they might be criticized as they spread misinformation and exploit the likenesses of people.
Like any technology, though it's all about how you utilize it. Creative and forward-thinking creators and marketers are already using deepfake technology -- in a responsible and ethical mannertoto develop new forms of art and tell fresh stories and enhance their video campaigns.
In this piece this article, we'll discuss some examples of the use of fake technology to good ends, as well as ideas about how to play around using the technology yourself.
What's a fake?
A deepfake is a video or audio record of someone whose facial or body has been digitally modified. Deepfakes make use of AI to create likeness, by using patterns to identify facial tones, expressions as well as movement.
Other terms for a deepfake can include synthetic or artificial media or AI-generated media.
An overview of deepfakes
The invention of the Generative Adversarial Networks (GAN) began the trend towards realistic fakes in 2014. GANs are made up of two AI agents which make fakes and recognize forgery, allowing the AI to improve over time.
They can also be made using a deep learning computer network known as a varial auto-encoder (VAE). VAEs are trained to encode images in low-dimensional representations of a object and decode the representations into motion.
The term "deepfake" wasn't coined until 2017 and, in the year, the media almost collectively alerted the public about deepfakes - with the first viral fake videos of Barack Obama and Donald Trump appearing across social media.
But deepfakes also have uses which are getting more useful to ordinary marketers and not only hackers or internet trolls trying to disseminate false information.
What are the effects of deepfakes?
Machine Learning AI is an essential part of developing a deepfake. Deepfakes use this technique to identify patterns in the visuals and data.
To create a fake deepfake video, a developer needs to feed these machine learning algorithms with long hours of real footage, which then trains the deep neural networks to identify the patterns in tone facial expressions, tone other things. Next, it is time to combine the learnings and graphics.
It doesn't take much to make a fake, simply existing videos or audio of who you're trying to replicate. And although it may seem difficult at first creating a fake doesn't require complicated tools - the only requirement is basic knowledge in graphic design as well as video editing abilities.
Some examples of artistic video deepfakes
Marketers are still in early stage of using deepfakes and other AI technology for video and digital marketing. These examples of fakes don't exactly fit into a marketing toolkit just yet however, they demonstrate the potential of these AI technologies in the moment.
1. Chris Shimojima's "Dolche Big Man" by Chris Shimojima"
The stunning Staff Picked music video from director Chris Shimojima takes deepfake technology and flips it over its head, using the faces of 14 performers (and 40 contributions) to tell a single tale. The result is a stunning, unexpected combination of tech and human expression.
2. David Beckham's numerous languages
Malaria Must Die employed AI to manipulate soccer star David Beckham to speak in nine different languages, leveraging deepfake technology in order to create a big splash and significantly increase the campaigns' reach.
3. Salvador Dali's museum greeting
It took nearly 1,000 hours of machine learning to allow the Dali museumMuseum to get their version of deep fake Salvador Dali exactly the way they want it. The new technology gives visitors to museums an entirely new view and allows them to gain knowledge about art directly from the creator himself!
3 everyday uses for deepfake technology for video
While some applications of deepfakes might be out of the average marketer's reach, there are many creative and innovative ways you can use the technology of deepfakes in your work.
- Repair sloppy lines in the post For anyone with any editing expertise, you know the trials and challenges of mixing polished audio from a more informal interview. Whether your subject misspoke or didn't respond with an entire sentence, using deepfake technology to fill in the blanks can be a fantastic method to keep your post production process moving with no need to reshoot. (Just be sure to get the subject's permission first, of course!)
- Customize videos for customers on a large size Marketers can use an easy way to personalize their video greetings or promotional videos with names of prospects as well as their names and company names. All you require is their names and enough audio from your on-camera talent for deepfake tech to add it into any video.
- localize your videos Deepfake technology introduces a new world of easy translation. Instead of relying on subtitles, AI can add spoken, translated audio or created using an audio file or the initial actor's voice.
Technology advances, and new possibilities
We don't know what the future of AI is going to be however it is a fact that deepfakes aren't going anywhere. Just as with other AI-powered technologies (chatGPT for instance? ) People who are willing to experiment with deepfakes while keeping their eyes open to the possibility of pitfalls, are well-equipped to succeed in the evolving technology of video.