THE NEW "REALITY" OF EUROPEAN POLITICS
What is real? What is reality? Questions that have preoccupied philosophers for millennia. But few of them would have anticipated the level of distortion of our perceived reality in today's world. The use of AI to create deepfakes has changed the game. Our political reality is being transformed as we speak.
AMECA is one of the world's most advanced humanoid robots, currently based at the Copernicus Science Centre in Warsaw. Her answers are based on OpenAI's artificial intelligence GPT. When asked about the potential impact of deepfakes on politics:
PROMISE OR PERIL
Artificial intelligence (AI) is reshaping the political landscape in Europe, bringing both unprecedented opportunities and significant challenges. As the technology evolves, its influence on democracy becomes more pronounced, manifesting in various forms from enhancing communication to posing ethical dilemmas. Through this project, we delve into the multifaceted impact of AI on democracy, examining both the promises it holds and the perils it presents.
What are AI Deepfakes?
Artificial intelligence, or AI, is technology that enables computers and machines to simulate human intelligence and problem-solving capabilities.
"Deep fake" means AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful.
FROM POLITCAL WEAPON TO OFFICAL AI REPRESENTATIVE
Dr. Dominika Harasimiuk, Ph.D. in law, is a research and teaching fellow at the Faculty of Political Science and International Studies at the University of Warsaw. One of her research areas is the regulatory and ethical aspects concerning policies for the development of artificial intelligence in the EU.
Dr. Harasimiuk sees the use of AI in political advertising as a development that has the potential to undermine trust in European democracies.
"Press conference in the Ministry of the State"
In April this year the Danish opposition party Dansk Folkeparti (Danish People’s Party) shared an AI-generated video of the Danish Prime Minister, Mette Frederiksen, on the social media X. The video shows the leader of Dansk Folkeparti reacting to a press conference where the Prime Minister announces that the government has chosen to abolish all national holidays - including Easter and Christmas. Instead the government wants to introduce the islamic holiday Eid al-Fitr.
According to Dansk Folkeparti the video was a satirical input in the debate on the government's decision to abolish the public holiday Store Bededag. But the video was received differently by political colleagues. Instead of laughing along, voices from all across the political spectrum criticized Dansk Folkeparti for sharing the video.
The prime minister herself stated that the video – along with another AI-generated campaign video of her – “is destructive for democracy” and puts the fundamental trust in society at risk.
The video even resulted in the Danish Minister of Culture announcing that he would look into the possibility of banning deep fake-videoes of politicians. While no legal steps have been taken yet, the Danish parliament met in the beginning of June to discuss the matter. The meeting ended in a wide political majority agreeing on a temporary code of conduct for the use of deep fakes in politics.
In the code of conduct the parties have agreed to refrain from creating deep fakes of politicians or political candidates without the consent of the person in question. They have also agreed to clearly declare when a video or other material has been created using deep fake-technology.
Dansk Folkeparti, the creators of the video of the Danish Prime Minister, are not a part of the agreement. They still claim that the video was “innocent and humoristic” and emphasize that the watermark in the corner of the video clearly disclosed that it was generated by AI.
The level of AI literacy among the general public is still relatively low, and in order to know that something like watermarks or even AI-generated content is something you need to have at least a basic knowledge of what this world looks like, and of course the younger generations are much more aware of this than the older ones. Harasimuik says: "Watermarking solves some parts of the problem, but certainly not all of it. The level of AI literacy is still not high enough for the public to be really aware and careful about what they see, to filter it. That's the major issue."
What do people think?
What do Poles think about the connection between artificial intelligence and political deep fakes? We asked Warsaw residents.
"Leaked emails"
In August 2023, the largest opposition party at the time, Poland's centrist Civic Platform, released an election ad containing real clips of former Prime Minister Mateusz Morawiecki, as well as text spoken by an AI-recorded voice of the PM reading supposedly leaked emails between the PM and his chief of staff. The content of the emails contradicted the PM's public statements of the proclaimed unity of the United Right.
Morowiecki's government admitted that some of the PM's emails had indeed been leaked, but subsequently denied that the alleged statements quoted in the video were true.
Dr Harasimiuk, as a Polish citizen, sees this certainly as very problematic. “For me, even if this is seen as legal, it is unethical and its unfair towards the voters.”
In this case AI was used to produce fake statements, of unknown origin to slander a political opponent. Furthermore, the publishers of the video did not state in any way that the content was distorted by AI. Leading some viewers of the video to believe their PM actually made those statements which aroused a controversy around the question of ethics in political communication.
IS THE EU READY?
The global AI market was valued at over €130 billion in 2023 and is expected to grow substantially by 2030, up to nearly €1.9 trillion. Private investment now accounts for most of the investment in AI.
The US is leading private investment in AI (€62.5 billion) in 2023, followed by China (€7.3 billion). The EU and the United Kingdom (UK) together attracted €9 billion worth of private investment in 2023.
Between 2018 and the third quarter of 2023, almost €32.5 billion was invested in EU AI companies, compared with more than €120 billion in US AI companies.
Resourse: European Parliamentary Research Service
The AI act - a step towards regulation
Nearly three years after its first proposal, the Artificial Intelligence Act (AI act) was approved by the European Council in May 2024. Parliament endorsed the act two months prior to the Council. With 523 votes in favour, 46 against and 49 abstentions, it was supported by a strong mandate.
Called ground-breaking and a landmark law, the AI act aims to support the development and uptake of safe and trustworthy AI systems that respect fundamental rights. Whilst regulating potentially harmful technologies, the act also acknowledges that AI can create a variety of benefits to society including better healthcare and cheaper and more sustainable energy.
According to the European Commissioner Thierry Breton, the act is “regulating as little as possible and as much as needed.”
As the world’s first apprehensive AI law, the act can set a global standard for AI regulation.
During the plenary debate on the act, co-rapporteur Dragos Tudorache said:
“The EU has delivered. We have linked the concept of artificial intelligence to the fundamental values that form the basis of our societies. However, much work lies ahead that goes beyond the AI Act itself. AI will push us to rethink the social contract at the heart of our democracies, our education models, labour markets, and the way we conduct warfare.”
The act classifies different systems of AI according to risk. The higher the risk of the AI systems, the stricter the rules. The most dangerous systems, such as cognitive behavioral manipulation, are classified as posing ‘unacceptable risk’ and will be banned from the EU. ‘High risk’ systems are regulated.
Deep fakes are also addressed in the act. While the act does not outlaw the use of deep fakes, it introduces a transparency obligation to ensure that end-users know that they are interacting with AI-generated content. With the new law, deep fakes therefore need to be clearly and distinguishably disclosed as artificially generated content.
We interviewed Zaza Dobordjginidze an expert on AI, a Neuroscientist and Journalist on the topic. Please read down below the interview.
Zaza Dobordjginidze talked about the ethical challenges of using AI-generated deepfake voices in political ads. He highlighted the alarming ease of creating realistic audio deepfakes and their potential to spread misinformation, incite hate, and manipulate public opinion. In the interview he stressed the need for robust regulations, including AI detection tools and legislative measures to combat these threats.
Additionally, he raised concerns about using AI as a political spokesperson due to the lack of accountability and human judgment.
"Gaming and politics"
Here's a simple one. Nigel Farage - leader of Reform UK, an opposition party in the UK - appears to be streaming himself playing Minecraft.
The graphics are very poor and almost everyone will notice the distortion of his facial features and therefore recognise the video as a deepfake.
Nevertheless, the video was made by an unknown TikTok user using an AI application that almost anyone can use today. Imagine the advances in technology in ten or even five years' time.
So far, we do not need to worry about videos like this having a relevant impact on politics, but as a Farage spokesman said, they are: "quite funny though".
IS IT ALWAYS ABOUT POLITICAL DEFAMATION?
Victoriia AI - Ukraine's Virtual Consular Voice
For the first time in history, the Ministry of Foreign Affairs of Ukraine has created a digital person using artificial intelligence technologies that will officially comment on consular information for the mass media.
The representative of the Ministry of Foreign Affairs for consular affairs, Victoria AI, is based on a real person - Ukrainian singer and influencer Rosalie Nombre, who agreed to participate in the project of the Ministry of Foreign Affairs and act as a prototype for the representative. The project team photographed and digitized her.
Victoria AI comments will appear on the official resources of the Ministry of Foreign Affairs of Ukraine, the website and pages in social networks, and will also be transmitted to journalists by the press service of the Ministry of Foreign Affairs of Ukraine.
The use of a digital person generated by artificial intelligence for consular commenting is primarily a saving of time and resources of the Ministry of Foreign Affairs. Real diplomats will be able to be more effective and focus on other tasks of providing assistance to citizens.
Dmytro Kuleba
(Minister of Foreign Affairs of Ukraine)
"Why?" That's Dr Harasimiuk's first question about this use of deepfakes. Of course, Ukraine is currently at war, the question of resources in general and human resources in particular is different from other nations. But on the other hand, Victoriia claims that all the words she speaks are "human generated". So, there are "hopefully" people behind it. But then the question arises, why do they not use a human to do the same thing? "I have some problems with understanding the real purpose. I have some concerns about the question of why. It is at least questionable".
What people are saying about Victoria AI ?
Reviews show a mixed attitude to the introduction of artificial intelligence to provide consular support and inform citizens. Some people worry about Victoria's actions, but part of the people are excited.
THE ROAD AHEAD
Artificial intelligence is here to stay. Whilst the use of artificial intelligence and specifically deep fakes in political communication is a source of concern to many, the AI act is a step in a regulatory direction introducing transparency requirements on deep fake-content. To the knowledge of Dr. Harasimiuk up until now there are no examples of regulation specifically addressing the question of AI in political advertising, and that's a problem.
Dr. Harasimiuk believes distrust can be a threat to European democracies. “It’s influencing political choices and it’s undermining the democratic process as such so the stake is really high and there is a fundamental risk to the integrity of the political processes.”
New and more advanced technology is constantly developing. In the coming years, the EU, national governments, companies and regular citizens will have to navigate a changing technological landscape. In that light, the AI act can be seen as the prologue of a story that has yet to be written. According to Dr. Harasimiuk, Europe will need further legislation and educational efforts to protect its democratic values.