There is a rising debate in academic and political circles regarding the relationship between artificial intelligence (AI) technologies and democracy. While some believe that there are expected negative impacts from using these technologies in the democratic sphere—such as distorting facts before the public and manipulating voters—others argue that AI enhances the future of democracies around the world by providing voters with databases about political representatives, combating corruption, and ensuring individual rights.
In this context, the Friedrich Naumann Foundation presented a research paper titled “Applications of Artificial Intelligence in Enhancing Democracy” in 2024. This paper aims to assess the anticipated benefits of using AI to promote democracy while not overlooking potential risks through real-world examples, and to draw general conclusions about the nature of the relationship between these two fields, as well as how to avoid the modern technology risks to individual freedoms and rights, which are fundamental to democracy.
Key Concepts:
Artificial intelligence is a technical system designed to perform specific tasks based on defined rules called algorithms. These algorithms are distinguished by their ability to evolve independently beyond pre-established human-defined rules, relying on the amount of training data available and computing capacity. Democracy, on the other hand, refers to a political system based on the rule of the people, through various forms of political participation such as referendums and elections. The concept of democracy concerns the sovereignty of several values: separation of powers, rule of law, protection of human rights, and citizen rights.
According to the research paper, opinions vary regarding the potential of AI to build democratic foundations. Proponents argue that AI could lead to an informational revolution, thus providing broader options for decision-makers while offering citizens promising opportunities to access information, stimulating their political participation. This means AI could contribute to achieving the goals of the democratic agenda by shaping future strategies related to voter needs and candidates’ views on issues such as climate change and infrastructure.
Conversely, opponents of AI usage cling to possible risks, such as the ability to distort facts and thus direct public opinion according to pre-set algorithms. Additionally, there is the concern of replacing humans with machines in jobs, ultimately concentrating wealth in the hands of a few within societies, which may impose their conditions on the world and undermine democracy.
Applications for Supporting Democracy:
The research paper highlights various actual applications of AI in the democratic field while emphasizing the need to acknowledge the threats posed by these applications, as follows:
Individual Rights: Ensuring individual rights is one of the fundamental demands of democracy, such as the right to social security, health insurance, fair working conditions, and equality regardless of gender or religion. AI applications can contribute to achieving these goals. For example, in the field of health rights, the (i-PROGNOSIS) app has been used in Europe since 2017 for early detection of Parkinson’s disease by being installed on smartphones to monitor voice quality or typing speed of users to identify the onset of the disease. However, this type of application carries risks of violating users’ personal freedoms, along with the potential for social assessments of citizens based on their behaviors, similar to the Chinese social credit system, which has been implemented in China since 2014 in various provinces.
Oversight Mechanisms: Monitoring the activities of authorities is one of the pillars of democracy to reduce the likelihood of financial and administrative corruption that misuses public funds without accountability. Here, Transparency International presented an anti-corruption program in Ukraine called (Dozorro) in 2018. This program utilizes big data on tenders and bids to develop a model capable of detecting tendering processes with a high likelihood of corruption. The risks here lie in the need for a massive database to ensure its effective operation, thereby requiring substantial financial allocations. Additionally, these applications may conflict with the principle of balancing privacy and transparency, raising the question of how much information about representatives should be tracked to ensure anti-corruption efforts while avoiding violations of their private lives.
Separation of Powers: Democracy cares about separating the executive, legislative, and judicial branches to ensure a balanced distribution of power and mutual oversight to limit potential corruption. The AI application (Codefy) is currently being tested in some German federal states to assist the judiciary by structuring and categorizing legal and commercial documents, with the ability to classify and analyze pleadings—thus saving judges time and effort and reducing errors. However, there are also risks associated with these applications concerning erroneous discriminatory outcomes that may arise from inaccurate or incomplete inputs or lack of data transparency. For instance, the (COMPAS) program in the United States assesses the probability of repeat offenses by convicted prisoners, and tracking these results revealed discriminatory outcomes against Black individuals.
Minority Rights: Protecting minority rights, linguistic diversity, and equality of opportunity are fundamental to democracy. In this area, the (Hua Ki’i) application aims to revive the use of rare languages like Hawaiian by processing various inputs such as images of objects and places, then providing translations of those objects’ names from English to the previous language. To ensure the app’s success, it is essential to provide dictionaries and grammar rules for both languages. Nevertheless, these applications do not guarantee the prevention of marginalization of minorities and their deprivation of rights; rather, laws and regulations must be enacted to guarantee minority rights, along with raising community awareness of the importance of integrating diverse elements into society.
Political Education and Information Verification: The concept of political education is a vital component of the democratic system, as it helps citizens understand political processes and access information easily and transparently to enable their participation. With the spread of misinformation, there was a need to dispel allegations and lies while working to identify accurate news. Therefore, the German Federal Ministry of Education and Research funded the (DeFaktS) application, which aims to combat and expose disinformation by extracting data from social media and messaging apps, subsequently analyzing it to identify the factors and methods linked to misleading information. However, there are risks that AI systems may produce unethical, illegal, or even discriminatory solutions; that is, the solutions might be beneficial to a specific group in society, necessitating evaluation of AI outputs based on legal and ethical considerations.
Key Conclusions:
The research paper reached several key conclusions regarding the relationship between democracy and artificial intelligence, the most prominent of which are:
- Various AI applications present promising opportunities to strengthen the foundations of democracy to varying degrees, but effective democratic structures are required to facilitate the exploitation of these applications. The availability of AI does not necessarily lead to strengthening democracy; this depends on the political authority that can use the applications to either enhance or undermine democracy if it chooses.
- Attention should not be limited to using AI applications that can generate new content and elements such as images, videos, texts, and sounds from simple text commands; these applications can be costly, and not all countries will be able to utilize them. Therefore, less complex applications should also be considered.
- AI systems benefit both citizens and officials alike, as the availability of massive datasets allows for better healthcare systems, and citizens can obtain personalized information about potential political representatives. This similarly applies to the state in various areas such as legislation and oversight.
- AI systems must be characterized by transparency and traceability to ensure their legitimate use and prevent manipulation, thus benefiting citizens and facilitating their search for reliable information while assisting decision-makers in planning, strategizing, developing systems, and making necessary adjustments.
- AI systems facilitate participation but do not necessarily empower participation for all groups, as the appropriate technological infrastructure may not naturally exist for all parties to enable their engagement. Relying solely on AI as a means to increase participation could harm technologically marginalized groups.
- The decisions and outputs of AI programs rely on the quality and type of input data; therefore, the current global digital transformation trend can be leveraged to provide accurate data. More importantly, suitable technological infrastructure must be provided alongside a trained human workforce capable of creating applications from the outset.
- Some advanced uses of AI fall under moral and legal responsibility, given AI’s capacity for self-evolution beyond human-defined rules; this raises questions about the influence of AI-generated recommendations on decision-makers and the potential for the exclusion of humans from critical positions.
In conclusion, the research paper indicates that strengthening democracy through AI applications requires the concerned states to provide suitable technological infrastructure and, crucially, regulatory and legal frameworks to ensure that increased technological advancement does not deprive individuals of their rights, which are the essence of democracy. As an initiative to prevent the replacement of humans by machines, the German Ethics Council (Ethikrat) calls for the inclusion of essential human elements when establishing evaluation criteria for AI systems, such as accountability, ethical disposition, ability to act, and rationality, to mitigate the risks these systems pose to democracy.
Source:
Nikolai Horn, Matthieu Binder, “How Technological Progress Can Strengthen Democratic Structures,” Friedrich Naumann Foundation for Freedom, April 2024.
