[column] Applications of deepfakes for brands and advertising

Kietzman and colleagues discuss what deepfakes are, how they work and what they can mean for advertising and brands. I share their key learnings here. The term deepfake is a combination of ‘deep learning’ and ‘fake’. It is a form of video (or audio) manipulation in which artificial intelligence creates a video that is fake, but super realistic. Deepfakes are within everyone’s reach because artificial intelligence is developing and software is easy to download. Consumers are getting creative with face-swapping videos like Avatarify and Reface, and creating avatars with FaceApp. With the Myheritage app you can bring old family photos to life. Who hasn’t played with the Wombo app yet?

Fake it till you make it
You create self-made influencers that perfectly match your brand and who spread your message through various platforms. Because you fully control what the deepfake influencer posts, you also have control over the content. Moreover, using deepfake influencers saves you a lot of money, they are available day and night and they do not cause scandals. By using deepfakes, you give consumers a leading role in your campaign, attract their attention and make it personally relevant. For example, consider a virtual clothing store that uses deepfake technology to swap a model’s face with that of the consumer. Deepfake images of people generated by AI are already for sale for professional purposes.

As a brand, you respond more effectively to customers and (international) markets, since you are no longer tied to one language or location. You let Beyoncé say something about your brand and contractually stipulate what she delivers for this and that she does not have to come to the studio for this. Also saves her time. Advertising agencies create new commercials by adapting original footage on demand, and in real-time in response to current events. This way you involve your target group as much as possible with your brand.

Real or fake?
The most interesting, but at the same time the most dubious feature of deepfakes is that consumers can hardly distinguish between real and fake. That makes them insecure. Insecure consumers are losing faith in brands. Even worse, if people continuously see expressions of your brand that could be fake, they will no longer trust what they read, see and hear about your brand. They then reach a state of reality apathy. Increasing consumers’ media literacy about deepfakes is an important agenda item for the coming years. For consumers to distinguish real and fake, for advertisers to prevent apathetic consumers.

The video below shows that deepfakes are becoming increasingly difficult to distinguish from reality.

Quality mark
Criminals misuse manipulated images for reputational damage and blackmail purposes. They commit identity theft through voice cloning. There is a need for regulation. Think of quality marks in commercial deepfake videos, so that people know that the messenger is a deepfake and has been developed with the permission of all parties involved. A digital fingerprint for real influencers distinguishes them from deepfake influencers. By watermarking every video created, just like branded and native content, you make it clear to consumers whether it is the original video or a modified version.

Deepfake ban
In early 2020, Facebook and Twitter decided to remove all deepfake videos that advertisers post on the platform that cause harm to others. The science examines the possibilities of deepfake detection, content authentication and deepfake prevention and the deployability and effectiveness of deepfakes for brands. Nevertheless, it is ultimately the advertising industry’s responsibility to use deepfakes in their campaigns in a constructive and transparent manner.

In short:
* Take deepfakes seriously, because they already exist.
* It is difficult to determine the value for advertising and brands, but keep an eye on this in the coming years.
* Deepfake use requires transparency in our industry and regulations. Make it clear to consumers whether it is an original video or a modified version.

Paul Ketelaar
senior Assistant Professor at the Behavioral Science Institute (BSI)
Radboud University

Thanks to knowledge base SWOCC. Image: Kulbir / Pexels

Related Articles

Back to top button