Even if you have never heard of the term ‘deepfake’ before, it is quite probable that you know what it is but did not know that this was the exact term that has been used to refer to it. In short, ‘deepfake’ is an amalgamation of the phrases ‘deep learning of Artificial Intelligence’ and ‘fake.’ What it means is that it is a method by which realistic fake content is created using AI.
Today, deepfake is no stranger in our digital world, appearing in our movies, audiovisual content and social media interactions. It is used to replace the voice, face, and even images of what we consume on our screens and devices.
In recent times, AI has been used to produce deepfakes of public or popular figures speaking about one side of controversial issues or international events that have been circulated on social media. When such views do not originate from the person in question, it becomes highly unethical and an infringement of that person’s rights and freedoms.
Deepfakes in the world of marketing
While deepfakes have been around for some time, we are hearing about them more often now because of our easy access to digital media and because much of our lives now centers around the online world.
A good example of deepfake being used in a movie is when the deceased actor Paul Walker made an appearance in Fast & the Furious 7 in 2013, with the technology being used to map his face to his brother’s body to bring the actor to life for the movie.
With deepfake technology thus being so useful, it has certainly made its way into product marketing in myriad ways. One of them is influencer marketing. You probably guessed how the tool is used here. Popular figures are superimposed into influencer marketing campaigns to either share an important message with a certain audience or to attract viewership for a certain niche market product.
These sorts of campaigns can be beneficial to audiences and are harmless in terms of ethics as the audience is not hoodwinked into believing something is real. There remains an unspoken understanding between campaign creators and their audiences that deepfakes, although fake, generate an authentic message.
Another useful manner in which deepfakes are used is when customers want to virtually purchase a product such as clothing. Here, an existing model is used to try on the clothes with the model’s face changed to the customer’s own, allowing the latter to make an informed decision before purchasing the item.
The downside of deepfakes
As more and more AI technologies are used to produce deepfakes for campaigns and online messages, it will become extremely challenging to differentiate between what is real and what is not. The danger obviously associated with it is the spread of wrong information and, as in the case of sensitive world issues, the usage of deepfakes to create disharmony and unrest.
In this vein, deepfakes can be used to impersonate public figures, employees of various companies, influencers and so on to spread misinformation.
Social media and its role in combating deepfakes
Experts suggest the development of algorithms on social media that can detect deepfakes and prevent their usage on social media platforms. To do this, social media companies need to cooperate in developing and adopting the relevant mechanisms to expose deepfakes and remove them from circulation.
Social media users too, have a responsibility in this that they are vigilant when consuming online material and follow campaigns that disseminate sensitive information. Fact-checking, verifying the authenticity of the information we consume, and refraining from downloading random attachments should be part of our social media usage practices.
Furthermore, regularly examining one’s social media privacy settings and placing restrictions on how much personal information one shares online are important to not fall prey to deepfakes.
On a larger scale, blockchain technology has been extremely useful in identifying deepfakes using ‘smart contracts’ that flag down unauthentic material to companies and users.
At the other end of the spectrum, Google’s SynthID has also been used with its ‘watermarks’ in audio content generated by AI. This makes it possible to demonstrate that the audio content has, in fact, been AI-generated, removing any malicious intent.
Overall, awareness about Deepfakes is important to enable the proper combatting of misinformation and unethical use. This involves businesses, cybersecurity companies, governments, and digital media users working together to overcome the threats posed by tools and technologies generating deepfakes.
(Anouk De Silva)