A WEB

-Unofficial summary by the A-WEB Secretariat who attended the Conference online- 

Conference ‘Artificial Intelligence, Democracy and Elections’

 - 50th anniversary of the CNE of Portugal-

ⅠSummary of the Panels

- AI-based news content raises concerns about reliability, and the spread of fake news may be exploited for election interference and political propaganda.

❍ Electoral System and the Introduction of AI

 ❍ Risks and Ethical Issues of AI in the Electoral Process

Conclusion of the Panel 1

AI can be used as a tool to increase the efficiency and transparency of elections. Still, it comes with risks, such as spreading misinformation, disinformation and manipulating public opinion. Therefore, maintaining a balance between the development of AI technology and the protection of democracy is crucial.

❍ AI, Election Campaigns, and Misinformation

Conclusion of the Panel 2

AI could enhance the fairness and transparency of democracy, but it also poses risks such as the spread of misinformation, manipulation of public opinion, and reinforcement of political bias.

It is important to balance the regulation of AI technology and the protection of democracy with ongoing legal and ethical discussions.

Adequate responses are required to ensure AI does not threaten democracy while maintaining a balance between freedom of expression and regulation of false information.

==================================================================================================​====​

II Detailed Information 

• Electoral management bodies (EMBs) and democratic institutions are facing numerous challenges, and the use of AI and its impact are important topics for discussion.
• Discussions on the ethical and legal limits of AI and the resulting risks and benefits are necessary as technology advances.
• This international conference is held with the support of the Speaker of the Portuguese Parliament, who emphasized the social importance of discussions on technological advancements while seeking ways to harness the potential of AI while adhering to democratic principles.
• The Portuguese Parliament is monitoring this issue, and hopes the conference will be a valuable experience for all participants.

❍ Eduardo Veracruz Pinto, Chair of the Department of Law at the University of Lisbon

• Electronic interventions such as algorithms and social media manipulation are threatening democracy during elections. While democracy is based on the rule of law, it cannot be defined solely as legal rules. Legal rules and legal norms are different, with norms being tools of governance and legal rules tools for achieving justice.
• To protect democracy, cooperation between academia, the legal community, and electoral management bodies is necessary; otherwise, the rule of law cannot be realized. In an era of advancing technologies like AI, democracy may be vulnerable, but it must overcome this vulnerability to become stronger.

❍ José Santos Cabral, President of the CNE of Portugal

• AI is fundamentally changing our lives by processing large amounts of data and analyzing patterns. However, without ethical considerations, it could threaten democracy.
• In the recent Romanian election (2024 presidential election), a conflict emerged between constitutional court rulings and manipulation of voter consciousness. A survey of 3,000 Europeans showed that two-thirds of respondents were concerned about the possibility of election results being manipulated by hackers.
• AI can become a tool to strengthen democracy, but if misused, it could threaten core democratic principles. Specifically, algorithms that prioritize information directly affect citizens' perceptions of reality.
• AI systems must prevent biases that could influence election results and ensure transparency and accountability. Oversight agencies are needed, and in 2024, the UN and the EU will set global norms and ethical standards for AI and propose strict requirements for high-risk AI systems.
• The future of democracy depends on the development of technology that aligns with democratic principles, preventing misuse and promoting public welfare.
• It is also crucial to help citizens understand the impact of technology on their decision-making through digital literacy training.

2. Panel 1: AI and Elections

❍ Carlos Fiolhais, (Faculty of Science and Technology, University of Coimbra)

• Bill Gates mentioned the importance of AI, stating, “AI has the potential to change the world as much as the internet,” and the significance of AI expanded with the release of OpenAI's ChatGPT, which gained 1 million users in just five days.
• The introduction of machinery during the late 18th and 19th-century industrial revolutions replaced human labor, promoted automation, and has led to similar patterns of progress to this day.
• Reflecting on the early history of computers, from Charles Babbage's automatic digital computer to the UNIAC computer and the invention of the transistor, the concept of a "home office" was first realized with computers. The development of digital technologies and AI has seamlessly integrated into the devices we use.
• AI has become a tool accessible to everyone, mimicking human intelligence, with its functionality easily understood through Alan Turing's "Imitation Game" in 1950.
• AI operates based on large-scale computer power and the internet, evolving by mimicking the neural networks of the human brain. AI has the ability to learn from existing human knowledge and create new works based on that knowledge.

❍ Sofia Casimiro, Professor of Law at the University of Lisbon

• AI can be applied in various contexts, having both positive and negative aspects. While AI can play a positive role in fighting hunger, treating diseases, and addressing climate change, it also poses risks related to misinformation, manipulation of opinion, and the spread of fake news.
• The legal community should respond by creating new laws to regulate AI's impact. Current legal norms fail to adequately address these challenges, so new rules are needed.
• AI’s impact is especially profound on democracy and elections, as it can influence voter perceptions and elections themselves. Therefore, legal restrictions and oversight mechanisms are essential.

         ❍  Luísa Meireles, Director of the Lusa News Agency

• In the development of media, there is a conflict between traditional and digital media. Traditional media relies on specific regulations and trustworthy sources, while digital media does not, making it easier for misinformation to spread. With the development of digital media, the accuracy of news has been diminishing as individuals share information on social media platforms. Additionally, algorithms on social media tend to reinforce specific political orientations, exacerbating social conflicts.

o    Fake news is spreading rapidly on digital platforms, particularly on social media like TikTok and Twitter (now X), where individuals with extreme political opinions spread content that quickly goes viral. Algorithms show more provocative content to users, meaning they are often exposed to distorted or inaccurate information. As a result, people frequently take fake news and misinformation seriously, accelerating social division.

o    The spread of misinformation becomes even more problematic when there is an economic or political agenda behind it. Misinformation is used to intervene in elections, promote political propaganda, and create social conflicts. 

o    When talking about traditional media, there are regulations on sources. In digital media, fact-checking is becoming increasingly difficult. Many news outlets provide information that is inaccurate or distorted, diminishing public trust in the news.

o    A lack of digital literacy greatly affects the spread of information. Today, many people consume news and information on social media without critically analyzing it, and as a result, misinformation is often accepted as fact.

o    There are many people who share content without verifying the information, which lowers the quality of news. Additionally, people are consuming information more rapidly, often judging articles based only on headlines without reading them properly.

o  Media must be independent and free for democracy to function properly. However, the independence of media is being threatened today, and journalists are facing restrictions on their freedom to investigate. This situation weakens the foundations of democracy, making it essential to ensure the free flow of information and fact-based reporting.

• The Venice Commission is an institution under the Council of Europe that aims to support the legal and institutional reforms of countries to uphold democracy, human rights, and the rule of law. Its main mission is to provide legal opinions on democratic institutions and fundamental rights, such as electoral laws, political rights, and constitutional justice, as well as to promote cooperation among European electoral bodies and organize international meetings and capacity-building training.

o The Venice Commission analyzes the impact of digital technologies and artificial intelligence (AI) on the electoral process, recognizing both the benefits and the concerns. Digital technology enhances the efficiency of elections, strengthens transparency, and provides opportunities for political participation. However, it also carries risks, such as election manipulation and distortion of voter expression.

o  Therefore, the Venice Commission has established principles to ensure the respect of human rights when using digital technologies and AI in elections. These principles include ensuring freedom of expression, protecting personal data, regulating election campaigns, and addressing cyber threats. Additionally, given the threat AI poses to the fairness of elections, electoral bodies are advised to carefully consider the introduction of AI and maintain a balance between traditional electoral processes and AI usage.

o When using AI, electoral bodies must ensure fairness, neutrality, and transparency, as there is a potential for biases related to race, gender, and political orientation.

❍ João Almeida, a member of the CNE Portugal

 The introduction of technology in Portugal's electoral system has led to improvements in efficiency and accessibility through the integration of digital tools and systems in the election process. These changes were driven by electoral law amendments, the introduction of electronic voting in 2019, and the COVID-19 pandemic.

• Key technological advancements include:

1. Introduction of an automatic voter registration system for overseas citizens
2. Introduction of mobile and absentee voting
3. Creation of a candidate registration and political finance management system
4. Introduction of digital authentication keys and citizen cards to enhance voting security

·  The adoption of digital systems has significantly increased voter registration, particularly among overseas citizens.

·  AI is being used to improve the efficiency of voter registration, electoral roll management, and monitoring misinformation and fake news. AI can also be applied to election result prediction tools and voter guidance services. The digital process aims to be inclusive of all citizens, ensuring transparency, security, and continuous updates, especially for vulnerable groups.

·   However, resistance to change and the ongoing education of all stakeholders remain significant challenges. Portugal's electoral system is complex and transient, emphasizing citizen-centered operations over professional systems. The government manages only central services related to electoral management and censuses, while local governments enjoy considerable independence. This structure creates difficulties in standardizing and regulating election procedures, particularly in defining electoral crimes legally.

·       There are concerns about the government imposing regulations on digital election services without sufficient discussion. The government's shift towards technical management of election procedures, instead of ensuring citizen and judicial participation, has raised concerns about the potential for manipulation of public opinion and election procedures.

·         The "truth" in elections is not an objective scientific fact but a subjective concept formed through social relations and common perceptions. Focusing only on discourse and actions can lead to misunderstandings, and a broader political perspective must be considered.

·   AI tools are drawing attention due to their potential malicious use in elections, particularly related to political campaigns and political finance. Although electoral bodies have the necessary legal authority, there is a lack of resources and authority in external administrative areas.

❍ Conclusion

• The role of AI in elections is increasingly prominent. While AI could improve transparency and fairness in elections, it also poses significant risks that must be mitigated through ethical and legal frameworks.

• Artificial intelligence (AI) is not merely intelligence but can be described as a complex computational process that handles vast amounts of information. There needs to be a discussion about whether it can replace human characteristics such as creativity and judgment.
• The potential legal and political challenges posed by AI’s capabilities should not be overlooked, especially its impact on democracy, including institutions like elections. AI technologies could influence elections in favor of certain candidates or political parties, raising concerns about biased algorithms and data manipulation. Additionally, AI could selectively present information to voters during campaigns or influence election outcomes through AI-based polls and analysis. This could undermine the legitimacy of elections, making legal safeguards essential.
• Agencies that oversee and regulate the impact of AI on election processes must cooperate. The Constitutional Court is a key legal body that assesses whether AI affects elections in a way that contradicts the Constitution. The Election Commission plays an important role in maintaining electoral fairness and addressing related technical and administrative issues. The Constitutional Court and the Election Commission must collaborate to quickly analyze AI's impact on elections and develop legal responses.
• Human intelligence can be described by three core elements: grammaticality (the ability to understand language and logic), social interaction (the ability to understand social norms), and judgment (the ability to make ethical decisions). Questions persist about whether AI can fully replace these abilities.
• There is a need for legal review and in-depth discussions on AI's impact on the Constitution and the electoral system.

     ❍ Susana Aires de Sousa, Professor of Law, University of Coimbra

• When considering the relationship between elections and AI, one of the first issues that arises is the connection to electoral crimes. Electoral crimes can be broadly categorized into three types: 1) violations related to candidate registration, 2) violations related to election campaigns, and 3) violations related to election procedures. While the law specifies about 150 types of violations, it does not sufficiently address new issues like false information triggered by AI.
• AI technologies such as deepfakes, social bots, and echo chambers (the biased circulation of information) have been identified as tools that manipulate voter behavior. These technologies can distort the fairness of election processes, influence voter choices, and even invalidate elections.
• Some existing electoral law provisions address coercion, fraudulent acts, and deceptive tactics (including fake news), but they are outdated and have limitations in addressing the false information created by AI. For example, criminal law does not explicitly define "fake news." Some scholars question whether false information related to elections should be subject to legal penalties. It is argued that criminal law should only intervene when the integrity of the election process is genuinely threatened, and its application should be limited to acts that actually harm democratic legitimacy. Criminalizing fake news in the election context requires a cautious approach, involving consideration of how to assign responsibility to various actors, such as digital platforms, political figures, and individuals.
• AI regulation is evolving, and Europe’s regulatory framework is already moving towards banning certain AI systems. However, the current regulatory framework may not fully address the spread of false information via AI, necessitating more refined legal responses.

      ❍ Jorge Bravo, Judge of the Supreme Court of Justice

• When analyzing the impact of IT companies on democracy, we need to consider the concept of "technofeudalism," where a small group of tech giants shape the political landscape without directly threatening democratic institutions. This trend is contributing to a shift toward post-democracy, where formal democratic institutions remain, but political participation is gradually weakened due to economic and corporate interests. Voter apathy, the growing influence of transnational organizations, and the decline of intermediary organizations like political parties and labor unions reflect this phenomenon.
• Furthermore, in the age of "post-truth," emotional appeals often outweigh objective facts, leading to the spread of false information and weakening trust in democratic institutions. Digital platforms accelerate the dissemination of this false information, sometimes manipulated by certain groups for specific purposes. Therefore, it is important to maintain a balance between protecting freedom of expression and preventing misinformation from threatening democracy.
• The European Union is responding to this issue by distinguishing between types of information manipulation. "Disinformation" refers to intentionally manipulated false information, "misinformation" refers to incorrect information due to a lack of fact-checking, and "malinformation" refers to information that, though factual, is distorted for a specific purpose. Legal and institutional responses must be developed to address the challenges of the digital age, with a focus on maintaining a balance between technological advancements and the protection of fundamental rights.

    ❍ Francisco Teixeira da Mota, Lawyer and columnist

• Research on the impact of AI on critical thinking has shown that people using AI tend to focus on reviewing AI outputs rather than engaging in independent thought. However, this effect can be mitigated with a higher level of education.
• Legal and ethical issues cannot be ignored. In the UK, discussions are underway about allowing AI to be used by journalists and creators without fair compensation. Issues like copyright protection and licensing have become significant global topics. While the US is approaching these issues through court rulings, discussions are still ongoing in Portugal.
• A survey conducted in Portugal regarding AI use in the media found that both the traditional newspaper Público and the independent investigative journalism podcast Fumassa mainly use AI for transcriptions, translations, and basic research. Both outlets remain skeptical about AI's reliability, as AI-generated content fails to reflect individual characteristics, and transparency policies for AI use vary across media outlets. Ultimately, AI remains a supplementary tool and does not play an active role in core content production.

     ❍ Gustavo Cardoso, Professor, Researcher at the University of Lisbon

• The central issue in today's discussion is not the technology itself but how people use and sometimes misuse it. Looking at the 2024 Portuguese parliamentary and European elections, it was observed that emotional content on social media spread much more widely than political content. In fact, only 2% of the 300,000 most shared posts contained political content, and even during the election period, political posts accounted for a very small proportion.
• On Facebook, political content was so scarce that one had to scroll through more than 300 posts to find any political posts. In contrast, some candidates made a strong presence on platforms like TikTok and Instagram, focusing on content that triggered public emotions. Some political parties, despite having relatively small supporter bases, effectively used social media to expand their influence, and the use of bot accounts became a subject of controversy. The key issue here is not simply gaining visibility but the actual impact of these strategies on voters.
• Meanwhile, during the Portuguese general election, false ads appeared on YouTube, triggering controversy regarding Prime Minister Antonio Costa and former Prime Minister José Sócrates. The use of paid ads to spread misinformation became a concern as it could mislead voters. However, while misinformation may influence voter opinions, it would not be appropriate to invalidate elections based solely on it. If misinformation were considered legitimate grounds to invalidate an election, it could threaten the legitimacy of many other elections as well.

4. Closing Ceremony

  ❍ Teresa Leal Coelho, a member of the CNE Portugal

• Reflecting on the key topics discussed during this conference, it has been a valuable opportunity to deeply understand the impact of artificial intelligence (AI) on the Constitution, fundamental rights, and the electoral process. AI provides opportunities to improve access to fundamental rights, such as in healthcare and education, but at the same time, it can lead to various risks, such as privacy issues, digital biases, and misinformation.
• In the election process, AI can contribute to the advancement of public opinion surveys, electronic voting, and election campaigns, but it also poses a risk of enabling the spread and manipulation of misinformation, which threatens the political public sphere. While current laws regulate election-related crimes, the spread of misinformation through AI makes legal responses even more difficult.
• The digital age has expanded freedom of expression but also affects individual safety. Therefore, regulation of digital communication should be carried out in a democratic manner and, where possible, through global cooperation. There is also the argument that human nature, rather than AI itself, is the bigger issue. AI is simply a tool, and its misuse ultimately stems from human actions, which is something we must recognize.
• Ultimately, misinformation creates a boomerang effect that undermines the credibility of those who spread it. The key issue is whether the public is developing the ability to identify false information or if it is becoming increasingly influenced by it. I believe this discussion must continue moving forward.