Disinformation campaigns pose a serious threat to democracies and trust in political institutions. Disinformation campaigns are nothing new and have been used for centuries by political, religious and military actors to manipulate public opinion, damage opponents or influence war strategies. However, in recent years in particular, with increasing digitalization and ever easier access to new technologies, these campaigns are on the rise.
Deepfakes have added a dangerous weapon to the arsenal of disinformation campaigns. Videos of politicians making statements that have never been made are appearing more and more frequently. The dissemination of such fake content can be used to sow doubts, create enemy stereotypes or even create political instability.
How are deepfakes used in disinformation campaigns?
Damage to reputation
A particularly common use of deepfakes is to defame political opponents. Opposing politicians or activists are placed in compromising situations in which they accept bribes or display immoral behavior. A well-known example comes from the USA: in 2020, a fake video of Democratic politician and incumbent President Joe Biden emerged in which he allegedly made racist statements. Although the video was debunked as a deepfake, it caused a stir and aroused mistrust on social media. In authoritarian regimes, such videos are used to discredit members of the opposition and destroy their reputation, causing the population to lose trust in them.
Elections
Another example is the targeted use of deepfakes to influence elections. The current presidential election campaign in the USA provides an unprecedented example of this. US megastar Taylor Swift has published a post on Instagram in which she announces her support for US presidential candidate Kamala Harris. Swift’s motivation for taking this step was AI-generated images that suggested Swift would support Donald Trump. Trump had shared these images on his social platform “Truth Social” to show Swift’s support for his election campaign. This example is an impressive illustration of the far-reaching consequences of disinformation.
In India, deepfake technology was used for the first time during an election campaign. The BJP (Bharatiya Janata Party) created a video in which a local politician, Manoj Tiwari, spoke in different languages to reach different groups of voters. This video used deepfake technology to make it look like the politician was fluent in the respective language. While this was not used as a disinformation campaign, it shows the potential of using deepfakes in election campaigns to manipulate voter groups.
Undermining trust in institutions
Deepfakes are also used to deliberately undermine trust in state institutions. This is done by discrediting politicians, government representatives or institutions themselves through manipulated content. One example of this is the use of deepfakes during the war in Ukraine, when a video of Ukrainian President Volodymyr Zelensky allegedly calling for surrender emerged. The video was quickly recognized as fake, but the attempt to undermine the moral stability of the Ukrainian people was clear. This illustrates how deepfakes can be used as psychological warfare.
How are such campaigns actually implemented?
The implementation of such campaigns requires a strategic combination of technological know-how, social manipulation and distribution via suitable channels.
- Creating deepfakes
The creation of a deepfake usually begins with the collection of sufficient image and video material of the target person. Machine learning methods, such as generative adversarial networks (GANs), are then used to create a realistic model of the person, which can later be inserted into any scenario. This is particularly dangerous if the target person is placed in an already incriminating context, for example in a confidential conversation or in a room with other questionable people. - Targeted distribution via social media
Once the deepfake has been created, the next step is to distribute it in a targeted manner. This is where social media plays a crucial role. Videos can achieve an enormous reach at high speed via platforms such as X (formerly Twitter), Facebook and YouTube. What is particularly perfidious is that such content is often disseminated by bots and fake accounts, giving the impression that it has been shared by real users. This creates what is known as a “viral effect”, where the video reaches millions of people within hours before it is possibly exposed as a fake. - Integration of fake news and narratives
In many cases, a deepfake is accompanied by a broader narrative strategy that is flanked by fake news and articles. For example, a video is released showing a politician in scandal, and at the same time fake news reports appear that further embellish this incident and provide additional details that further undermine public trust. These reports are often published on websites and blogs that appear to be trustworthy but are actually part of an orchestrated disinformation network. - Targeting specific groups
One of the greatest strengths of modern disinformation campaigns is the ability to target specific groups. With modern algorithms, deepfake videos can be shown to exactly those people who are most likely to fall for the fake. For example, videos targeting a specific political party could be sent to their voter groups in order to unsettle them and influence their opinion.
The purpose behind the campaigns
The target of disinformation campaigns with deepfakes is diverse and can vary depending on the political or military context:
- Dividing society: Creating a deep mistrust of media and political institutions to encourage social unrest and division.
- Weakening democracies: Democracies are often about unsettling the population and undermining trust in the electoral process and the legitimacy of politicians.
- Psychological warfare: In conflict situations, deepfakes are used to create confusion, break morale and destabilize the enemy.
Sources:
