In the recent past multiple years the development of the deepfake as a technology transformed it into the powerful means that brings high risks to political and democratic processes. Although deepfakes (artificial intelligence-based videos, images, or voice recordings, which persuasively impersonate real individuals) may be put to entertainment or satirical purposes, their application in the political field is becoming even more troublesome. Although we might once have associated deepfakes with YouTube and realistic music videos, in politics the most outrageous hoaxes are deepfakes: fake speeches, synthetic scandals, and other fake political news.
This blog post details what political deepfakes are, how they are created, real-life instances as well as what measures societies can implement as a solution to this ever-enlarging threat.
What are Political Deepfakes?
Political deepfakes are artificial media produced by applying artificial intelligence (deep Learning models) in order to recreate the appearance of political actors that state or do things they never said or did. They may include:
- An artificial clip of a candidate supporting protesting opinions
- A manipulated speech that will encourage furor
- Recording of a leader uttering inflammatory utterances
- Fakery of an interview or press conference
Due to their high similarity to the original image, it is very unlikely that viewers can perceive deepfake as a fake, which is especially true when it is disseminated lighting-fast without verification in social media.
See also: How to Create a Successful Photo Portfolio?
What is the Process of Political Deepfakes?
The process of creating a deepfake includes machine learning but, mostly, Generative Adversarial Networks (GANs). Those systems are taught by looking at hundreds or thousands of pictures or video clips of someone to figure out that individuals facial expressions, voice patterns and mannerism. After training the model is able to produce convincing fake content that resembles the appearance and voice of the person.
A deepfake can be produced by virtually anybody with elementary technical skills with sufficient data and energy sources to implement them. This easiness turns deepfakes into a threat in the political realm where a couple of seconds of fake videos can shape opinion.
Why are the political deepfakes so threatening?
The aftermaths of political deepfakes can be very long:
1. Disinformation Campaigns
Deepfakes allow sharing wrong information and propaganda when it comes to elections, referendums, or conflicts between countries. A viral misinformation is difficult to un-do.
2. Voter Manipulation
Entrances of fake footage of the candidates saying racist, sexist or inflaming words can change the course of the voters, in particular, those who have not decided to vote.
3. Deterioration of trustfulness
The appearance of deepfakes into the consciousness of people can lead to their becoming suspect in all media texts. This results in what is called the dividend of lying, i.e. the factual situations may be discredited as being feigned and nothing is taken as credible.
4. Diplomatic Tensions
False pretenses of political rhetoric or military infections might include strain amid the countries to an extent of practically coming into a war.
Political Deepfakes in real life
Not all deepfakes have directly influenced politics, but nonetheless, there are a number of well-known cases that could demonstrate the danger:
Ukraine-Russia Conflict (2022): A deep fake video has emerged on the internet of the Ukrainian president Volodymyr Zelenskyy allegedly asking his soldiers to surrender. This video was soon disproved but it pointed to the use of deepfakes as a tool of war.
Barack Obama Deepfake (2018): In response to the vulnerability of misinformation being spread via technology, filmmaker Jordan Peele made a deepfake video of former President Obama; the point was to show how technology could be utilized to create lies. Despite being educational in intention, the video demonstrated how easy forging of political messages can be.
Indian Elections (2020): Deepfake of a politician speaking multiple languages has been widely disseminated by a political party to attract the voters of different groups. Though not ill-intentioned, the event triggered ethical questions of artificially personalizing political messages with the help of AI.
The Role of News Outlets and the Social Media
The most popular platform to download deepfakes is social media. Their algorithm based character allows spreading fake material in a short period of time, in case it involves outrage or strong emotions.
The old news channels, on the other hand, are coming under enormous pressure of authenticity of the videos, according to the demands of verification before the release of such videos. One error, including a mistake of broadcasting a fake video as real, can destroy credibility and become a misinformation factor.
As a countermeasure, others have implemented AI-powered detection mechanisms, and some are collaborating with fact-checking organizations to readily identify deepfakes. Nevertheless, the detection technology cannot always be as advanced as new deepfakes.
What is the Solution to Political Deepfakes?
The technological, regulatory, media literacy, and public awareness solution is a multi-layered attempt to combat political deepfakes:
1. Deepfake-detecting Tools
Tech companies and researchers are coming up with deepfake detection tools to find the deepfakes by inconsistencies in blinking, facial shadows, syncing of lips and the metadata. Although useful, these tools require frequent updates to stay in line with the advancing AI methods.
2. Regulation and Legislation
Certain countries are coming up with legislation to label intentional deepface production as a criminal activity. As an illustration, the DEEPFAKES Accountability Act in the United States suggests to label synthetic media. Legal systems are however very different in jurisdictions.
3. Media Literacy Campaigns
There is a need to sensitize the population on how they can detect fake news and videos. Skepticism toward the viral content and critical thinking skills can decrease the influence of deepfakes.
4. Sectoral Collaboration
Governments, along with the tech companies, journalists, and civil society organizations should collaborate to create the tools, disseminate the information, and act rapidly in response to political deepfakes becoming viral.
Conclusion
Political deep fakes are not the only technological curiosity, but the explicit and imminent threat to democracy. In a world where the credibility of the institution is already bad, the capacity to convincingly speak or act like a politician can be catastrophic in building a discourse in the society, election, or even international relations.
In order to save the democracy of societies, it is critical to ensure that societies understand the danger of political deepfakes and realize that it is imperative to respond to the same in time. The battle with deepfakes is also not the one related to technologies exclusively but the one aimed at safeguarding the truth itself.
