Smart elections or rigged algorithms: the rise of artificial intelligence in electoral governance in Southeast Asia

Syafhendry, Syafhendry; Ganaie, Nasir Ahmad; Yama, Arifeen

doi:10.3389/fpos.2025.1672310

ORIGINAL RESEARCH article

Front. Polit. Sci., 22 October 2025

Sec. Comparative Governance

Volume 7 - 2025 | https://doi.org/10.3389/fpos.2025.1672310

This article is part of the Research TopicPragmatism and Democratic Backsliding: Local Election Dynamics in Southeast AsiaView all 4 articles

Smart elections or rigged algorithms: the rise of artificial intelligence in electoral governance in Southeast Asia

Syafhendry Syafhendry¹^*

Nasir Ahmad Ganaie²

Arifeen Yama³

¹Department of Government Studies, Universitas Islam Riau, Riau, Indonesia
²Department of Political Science, University of Kashmir, Srinagar, India
³Institute of Asian Studies, Chulalongkorn University, Bangkok, Thailand

Introduction: Introducing Artificial Intelligence AI into electoral-management infrastructure has radically transformed the administrative practice in Southeast Asian democracies through streamlining voter registration, consolidating identification, and making the real-time observation of the electoral events possible. Although the benefits of AI are undoubtedly tangible in terms of efficiency, transparency, and accessibility, its appearance in contexts where such regulatory sanitation is antithetical presents sharp postulations of algorithm bias, lack of data security, and possible undermining of democratic values. The current study questions both positive and negative effects of the implementation of AI in Indonesia, Thailand, Philippines, Myanmar, and contrasts their benefits and limitations concerning the fairness of electoral processes and good governance.

Methods: The investigation methodologically includes the secondary sources, i.e., government reports, institutional white papers and NGO assessments from the 2019 to 2024 election cycles. It measures the degree of AI implementation, electoral procedures, and regulatory models within the individual countries using a comparative qualitative approach.

Results: The results indicate that voter verification and result-monitoring systems aided by AI amplified the level of administrative coordination in Thailand. However, it also created a sense of disquiet about the lack of explanations for unexplained anomalies in data and a high level of algorithmic transparency obscurity. Indonesia's strong and most advanced biometric voter identification system, the Philippines' cyber-based registration and results reporting strengthened the uniformity of procedures, but inconsistent security practices and anonymised analytic operations introduced doubt to goal-targeted campaign communication. The first use of biometric identification tools assessed in Myanmar highlighted the potential of biometric identification to create more easily accurate voter lists, even though infrastructure constraints and periodic system failures hindered the full implementation of biometric identification across Myanmar. As a cross-case synthesis illustrates, purely technical protection and an opaque political-economic commission provide no safeguards against the inadvertent amplification of existing electoral weaknesses and concurrently increase operational capability.

Discussion and conclusion: Finally, the study benefits the field of inquiry by providing empirically based knowledge about the real-world impacts and ethical considerations of using AI in developing democracies and suggests policy guidelines to ensure responsible and accountable AI incorporation.

Introduction

Due to the variety of authorities in Southeast Asia, the electoral systems found there are also very diverse. In Indonesia, the Philippines and Thailand, elections are held regularly, free of charge and competitively, in order to affirm the legitimacy of governments selected by people voting (Morgenbesser and Pepinsky, 2019). However, in places like Myanmar and Cambodia, where authority comes from the center, elections may be changed to achieve and maintain control for a dictatorial group. Exploring the way Artificial Intelligence (AI) affects voter practices and results in the region would be more meaningful if we realize that the region comprises democratic and non-democratic regimes (Tapsell, 2021). Throughout the last decade, more AI and digital solutions have been added to election systems, making it easier to run, monitor and oversee elections. AI is now part of many areas in election administration, such as signing up voters, monitoring votes as they happen and studying how people vote. Experts tell us these innovations can address persistent issues in election management, including fraud, problems verifying voters and late reporting of results (Yang, 2023).

Nevertheless, when more of our elections depend on AI, serious concerns about how it may be misused come to mind. Now that AI methods are being used more in election tasks, people wonder if they might affect the results with targeted information, misleading news and unintentionally biased results (Thomas and Andrew, 2020). The main research question for this study is whether the use of AI improves how elections are run and strengthens their reliability, or if it is exploited to set election results selectively. Endeavoring to analyse whether AI benefits or threatens elections in Southeast Asia, this paper looks at the pros and cons of its use (Tan, 2020). The study's value is found in how it examines the role that AI plays in Southeast Asia's elections, a region where politics are unstable, authoritarian forces are common and digital campaigns are widespread (Callahan, 2018). Because AI plays a key role in today's politics and is being used more, we need to know how it influences elections. In this study, we hope to help build a better understanding of AI governance, greater integrity in voting processes and safer online environments for democracy (Jayakumar et al., 2020).

Understanding the role of Artificial Intelligence in electoral management through review of literature

Electoral management is now much different, thanks to Artificial Intelligence (AI). Across the world, AI is beginning to support organizations with election management, to guarantee transparency and to look after people's voting data. Estonia is a pioneer in online voting, since it was first implemented in 2005 (Goh and Soon, 2019). Many in Estonia are now able to vote from home using special internet codes for their ID. Estonia's success with its system makes it a known example of e-democracy, and AI now plays a big part in making voter registration and identity verification secure and helps detect attempts at election fraud (Abdullah, 2017). In other Western contexts, AI technologies are similarly employed to optimize election processes. For example, People designing American elections rely on AI to study the likely participants, predict results and follow vote counting as it happens (Arifianto et al., 2019). With the help of AI, voters can sign up more easily, are encouraged to vote, and more precise predictions are made about the outcome of elections from gathered demographic information (Warburton, 2020). Election results and voting patterns in the United Kingdom are analyzed with AI, which speeds up the way information is reported. Even so, while AI can improve elections, using it in the process has aroused worry about transparency, accountability and security (Morgenbesser, 2017). However, AI is still new to Southeast Asia, even as the region moves toward using digital tools in organizing elections. While Indonesia and the Philippines have put in place new digital systems for voting and identification, the adoption of AI at a large scale is held back by problems with infrastructure and rules. Indonesia has put biometric voter identification in place, despite continuing controversy over how safe and reliable such systems are, with problems over privacy and cybersecurity often raised (Sinpeng, 2020). The Philippines provides online options for registering to vote and sharing election results, but many have doubts about their reliability, since they might be manipulated (Choi, 2019).

The dark side of AI: ethical challenges and risks of voter manipulation in the digital age

AI has the ability to influence how people vote, so people are worried about the effects on elections. Targeting political campaign messages with AI is possible thanks to the analysis of voter data, which can be used to tailor ads to each person's unique interests, demographics and actions (Suryani, 2015). Despite being called practical tools for involving citizens in elections, such ways to address individuals can also be risky for voters. An excellent case of AI groups trying to guide elections comes from the Cambridge Analytica scandal from the 2016 United States Presidential Election (Gomez and Ramcharan, 2022b). Analyzing Facebook data from many users, Cambridge Analytica used AI to target every voter personally with political ads meant to influence them. Thanks to this case, it's clear how greatly AI-led political strategies can affect democracy, which means that using individual data in politics raises important ethical issues (Courtin, 2023). Similar problems appear in Southeast Asia, since social media is increasingly used there to change political opinions by using AI technologies. There, Facebook plays a major role in people talking about politics, yet it has been criticized for providing a way for false news and focused political advertising to spread. These efforts can cause real problems for fairness and the quality of elections (Rappa, 2023a). Also, the rising worry is that AI systems give a big boost to divisive and radical information online (Segal et al., 2020). In both the 2015 and 2020 general elections in Myanmar, Facebook was a way for misinformation to spread, thanks to AI algorithms boosting hate speech and telling false stories about politicians and ethnic minorities. The recent use of AI-supported false information to influence how the public feels can seriously harm the trust and fairness of elections when that wrong information is directed at people who are easily influenced (Gomez and Ramcharan, 2022c).

Algorithmic bias, cybersecurity, and the emerging threats to electoral integrity

Using AI in managing elections brings a number of complex problems and important risks. One especially important problem is called algorithmic bias. If problems of bias exist in the data used to train AI, the systems are just as susceptible to them, causing algorithms to worsen ongoing social, economic or political issues. Algorithms in elections may cause workers to improperly register people from communities and may drive advertising that targets select populations (Tawakkal et al., 2017). In Southeast Asia, this issue is especially important due to the frequent connections between political and social differences and the way votes are cast, as well as the possibility that information about some groups is either missing or inaccurate in the data. One more serious risk in AI for elections is that these technologies can easily be attacked online and broken into (Dettman, 2023). Those systems that assist with voter registration and voting by email are at risk of being compromised by actors from abroad as well as local groups. Various cyber-attacks and accusations of bogus election interference in Southeast Asia have made digital voting security a top concern in the region. Earlier elections in Indonesia have suffered attacks trying to corrupt the election and data system, which has drawn attention to possible vote rigging (Grömping, 2018). In the Philippines, safety issues have arisen again and again using automated voting machines, and people have accused the country of tampering with ballot results during the 2016 presidential election (Fadillah, 2019). Besides, AI-generated deepfake videos now create extra risks for the electoral system. Such manipulated information can lead some members of the public to believe false stories about political figures. The rising skill of deepfake technology makes it more dangerous to hold elections, since it is now easier for someone to spread lies and often harder to spot them (Ricks, 2019).

Regulatory and ethical issues: transparency, privacy, and fairness in AI-powered elections

Plenty of ethical and regulatory issues attach to AI in the field of elections. Many people are worried about how artificial intelligence works behind the scenes. It is often impossible to understand the choices made by AI systems that help decide elections (Xia et al., 2019). Because we can't see how algorithms work, it is challenging to assess both their responsibility and effect on election outcomes. Since transparency in AI matters most in democracies for trusting their elections, it should become a priority to achieve transparency in these systems (Jaffrelot, 1999). AI and elections open up many difficulties about preserving privacy (Saxena, 2021). Many AI-centered political campaigns use a lot of personal data to customize the way they target voters. Though this practice might bring more people to vote, it worries privacy experts because it could expose people's personal info to those in politics. Because data protection is weaker in Southeast Asia than in Europe or North America, the dangers of misusing personal information are particularly clear there (Loughlin, 2018). Android Systems has always been in disagreement regarding fairness. Algorithms should be built so that no group is treated differently by them. Even so, strong laws and rules are missing, which can unintentionally increase inequalities by ignoring equal representation in elections. Separate government teams are needed to oversee AI and help create rules that guide AI use in ways that respect democratic values rather than rushing toward what is most convenient from a technology standpoint (Tapseii, 2018).

Emerging AI adoption in Southeast Asia: contextual challenges and the path toward responsible governance

Using AI in electoral management in Southeast Asia is taking shape, though adoption is still facing a number of issues. Biometric voter identification has allowed Thailand to noticeably improve how transparent its elections are. However, political upheavals and the military's frequent involvement in politics worry some about the chance that AI may be abused in elections (Pradityo, 2022). In Myanmar, too, AI is not commonly used for election monitoring since the country's political problems make it harder to use digital systems for holding elections (Simandjuntak, 2021). If AI is put to use for monitoring the public in Myanmar, it will raise big questions about the country's civil and political rights (Gonschorek, 2021). AI and digital solutions are now used in the Philippines to carry out electronic voting and to make result reporting real-time. Still, such technologies have been abused for politics through the use of targeted political ads and spreading fake stories. Because of AI-supported ways to influence voters, many people worry that recent elections have not been fair, as the experiences from the 2016 presidential contest prove (Rappa, 2023b). Different adoption of AI in Southeast Asian countries shows why it is necessary for each country's regulations to be designed for their particular problems. A way needs to be found to control the risks from voter manipulation, algorithmic flaws and safety at polls when using AI to increase election smoothness (Hicken and Tan, 2020). This literature review explains the complex connections between AI, electoral management and ethics, showing that having robust supervision is necessary to prevent AI from defeating democratic principles in Southeast Asia (Amick et al., 2022).

Research methodology

Our study continues to a large extent in the form of secondary entanglements-academic books, governmental reports, authoritative statistical summaries, and other credible publications and thus it allows a strictly thorough study of the workings of Artificial Intelligence (AI) in the field of electoral administration. The goal of this research is to use data to better examine how AI technologies impact elections, voting and how people perceive government. This research utilized the use of secondary data, which comes from government sources, statistics on elections, recognized social media data and surveys of public opinion. They offer plenty of information on AI usage in elections, allowing the study to examine these practices in a number of countries in Southeast Asia. Reports by national election commissions give us important facts on voting, the election outcomes and how technologies like biometric systems and online monitoring systems were used. With these data, we learn how AI is included in elections and if its use causes noticeable enhancements in efficiency, transparency and how reliable elections are. Apart from what is available from the government, Facebook, Twitter, and Instagram offer a lot of information about political advertising powered by artificial intelligence. Metrics looked at by the study include spending on political advertisements, involvement by voters and the type of voters being reached. The study examined how artificial intelligence changes political messages, examined its effect on voting behavior and addressed the challenges involved with using personal data to achieve this. Studies of public feeling about AI in voting will reveal the perspectives of the general public. Surveys, which the Pew Research Centre and Ipsos sometimes conduct, are used to determine how the public feels about AI in elections. Points of discussion will be such things as public trust in AI used for elections, concerns surrounding privacy and thoughts on whether AI in voting is fair. The findings from this data played a key role in recognizing the ethical and regulatory issues that come with using AI in elections, when things like voter manipulation or algorithmic bias are concerned. File findings will be studied using a combination of descriptive and inferential statistical approaches. Statistics will help show the main features of the data, including how many people vote, how AI is used in elections and how much trust the public places in computers managing elections. Core variables will be studied using measures of central tendency and dispersion, and correlations between AI use and electoral outcomes will reveal whether AI helps improve processes, get more people to vote or increase transparency. Comparative analysis by studying relevant published data has also been used to compare AI applications in electoral management among Southeast Asian nations. The study scrutinizes countries such as the Philippines, Thailand, Indonesia and Myanmar to examine how much AI is used and how that affects election results. By comparing different uses of AI, this approach will identify successful practices and reveal potential problems connected to its use during political and electoral activities. The current research is wholly reliant on existing sources, the credibility of which is predetermined by the substantiating, completeness, and soundness of information presented in them. Details about AI in elections can be either limited, extremely old or found in conflicting formats in various regions. Moreover, understanding perceptions from data is influenced by the way the survey was made and the respondents' bias, especially in places where AI in voting has not been well recognized yet. Still, the paper relies on a thorough process that uses data to analyze how AI is being used in electoral management in Southeast Asia. The analysis in this study relies on multiple data sources that explore how AI affects election results, public behavior and how people trust political groups.

Results and discussions

Examining the role of AI in electoral management in Southeast Asia

Artificial Intelligence (AI) is growing in its role in managing elections across Southeast Asia. AI is helping to modernize voter registration structures and keep an eye on results in real time. Even so, such developments lead to important legal and ethical questions, as they could make voting fairness a concern (Case, 2005). Here, this section explores the many tasks AI performs in Southeast Asia's electoral processes through voter registration and identification, monitoring elections and involvement in digital political campaigns, by featuring case studies from the region (Neher, 1994). AI has done much to smooth the system for handling voter registration and identification in Southeast Asia. Previously, both signing up new voters and confirming their IDs have created several challenges and suspicions. Because of these issues, nations throughout Southeast Asia are using AI to ensure that voter registration is both safe and accurate (Aspinall et al., 2022). In Indonesia, biometric technology is being used by the government to stop voter fraud and verify that every person is listed in the system just once (Tapseii, 2018). AI algorithms are used together with facial recognition and fingerprint scanning to check who voters are. Every biometric sample from a voter is compared on the system against a central database, so any issues are highlighted. Because they brought these AI technologies into voter registration, and due to this new mechanism, fraudulent voting in Indonesia decreased and electoral records turned out to be more accurate (Eva-Lotta and Andreas, 2010). A number of people worry about how secure such data is within a country where privacy laws are not fully developed. Just as in Thailand, the Philippines has adopted AI-powered fingerprint scanning for biometric voter registration (Gomez and Ramcharan, 2022a). The COMELEC launched the new system to manage fake or duplicate registrations. AI technology helps check and match voter biometric data against government records to stop anyone from registering fraudulently. People often praise biometric technology for reducing vote fraud, but the Philippines has encountered problems during its use, such as fears that both biometric data and the electronic system might be manipulated (Fitriana et al., 2023). As a result, AI is improving how and who is registered and checked during elections. Yet, because so many use it, reliable protections are required to keep private data from being abused. Table 1 below shows a comparative overview of AI adoption in Southeast Asian countries.

Table 1

Table 1. Comparative overview of AI adoption, regime types, and recent electoral cycles in selected Southeast Asian Countries.

The Comparative study presented in data Table 1 offers a methodologically sound and empirically strong grounded AI implementation in four Southeast Asian countries: Indonesia, the Philippines, Thailand and Myanmar, where the usage of the AI in electoral management has gained tremendous momentum. Indonesia demonstrates the usage of the most advanced and modernized form of AI with more security and multiple levels of authenticity, primarily through its usage of the biometric e-KTP initiative that reflects institutional supremacy and upgrades. The 2024 elections in Indonesia became a game-changer with the successful integration of these technological interventions. The Philippines' revelation of moderate adoption of AI technology, particularly in the new voter registration process during their 2022 elections, became a game-changer in their electoral management, which was targeted at reforming the electoral system of the country. Thailand's hybrid management reflects vigilant AI implementation, mainly oriented around electoral supervising and predictive analytics, as seen in the 2019 elections. The moderated AI adoption highlights regulatory ambivalence amidst political complexities.

On the other hand, Myanmar's authoritarian regime corresponds with minimal AI integration, restricted to surveillance functions during the 2020 elections, exposing systemic challenges to electoral transparency and democratic legitimacy. This comparative matrix explicates the critical interaction between regime characteristics, technological capacity, and institutional governance systems that shape and transform AI's multi-layered role in electoral processes within Southeast Asia. The scrutiny is validated by contemporary studies and policy reports from authoritative bodies such as International IDEA and RSIS, ensuring academic consistency and meticulousness.

Using data to forecast election results

AI is widely used in tracking elections and predicting results as they occur. Watching elections develop and combining that with artificial intelligence can improve transparency and reduce improper actions (Yusianto et al., 2023). It has been usual to do election monitoring with manual methods and on-the-spot observation, but this can be slower and error-prone. Nonetheless, working with AI makes the whole process easier and faster. Southeast Asia is seeing more use of AI algorithms to watch over elections as they happen, studying several data points to find signs of something unusual (Abbott, 2011). The systems are used to monitor votes, check online discussions and find reports about illegal activities during an election. As an illustration, AI technology was used during the 2019 Indonesian presidential election to spot any unusual behavior from voters, such as lower numbers of people voting in some locations. Monitoring large amounts of data on the spot, AI can alert election supervisors to possible electoral problems, letting them take action early (Teo, 2022). In addition, AI is used in Southeast Asia to estimate how people might vote and who might win an election. Experts have started forecasting election outcomes using tools that study previous voting, analyse people's ages and follow social media exchanges (Satrio, 2019). Even though these models help us understand voter trends, people are still worried about how accurate their predictions are and if they could be biased. AI helps experts in the Philippines estimate how many people will cast their votes and how candidates are likely to do. Even so, using such predictions to send targeted messages to voters can change people's opinions in politics (Ufen, 2023). Like many other AI systems, AI used in election monitoring and result prediction face serious ethical obstacles associated with accuracy and the likelihood of data being altered. Electoral process trust is best preserved by ensuring that the deployment and use of AI are completely clear to everyone (Schafferer, 2017).

Using social media platforms for election campaigns

Using digital platforms powered by AI has become an important part of election campaigning in recent years in Southeast Asia. Instead of reaching everyone, political parties use AI to offer personal messages to people whose data they have collected on social media like Facebook, Twitter and Instagram (Hicken, 2022). By offering such platforms, campaigns are able to customize what they say to each group of potential voters using ads and messages tuned to their interests, behaviors and concerns (Allen, 2017). It is most noticeable in the Philippines that AI supports online campaigning on social media. Politicians and parties use AI-based voter data research more often in the Philippines to place targeted ads for voters. Thanks to social media algorithms that use lots of data about internet use, campaigns are able to create highly individualized messages. Although focusing on likely voters inspires them to vote, it has caused people to worry about misleading the public (Case, 1996). Using AI, it is easier than ever for untrue information and charged political stories to find their way to social media, making it hard to verify what users see. AI-based campaigning has come under attack by some, who complain that it threatens democratic principles. According to critics, using AI to target campaigns tends to separate people further by giving voters more of the same political information and little chance for honest debate. AI's power to focus on individual voters using their personal details raises questions regarding what is right about privacy and how much effect it may have on voters (Washida, 2023). The data reflected in Table 2 clearly shows the comparative analysis of the AI adoption and electoral governance system in the studied Southeast Asian countries.

Table 2

Table 2. Comparative analysis of AI adoption and electoral governance in Southeast Asia.

Indonesia shows a high level of integration of AI, such as biometric voter-identification systems (e.g. e-KTP), combined with facial-recognition algorithms. Strong governance is essential for good public service (Rassanjani and Meesonk, 2025). While this has been subtle, vital identification measures have ensured stricter voter verification procedures and reduced electoral fraud. This success can be aligned with recently implemented legal institutions and courts overseeing the use of AI in elections. However, the hurdles are ongoing, predominantly relating to the ongoing application of statutory provisions and the development of safeguards for individual privacy rights. In the case of the Philippines, the application of AI is moving at an incredible speed, mainly in the form of social media surveillance for detecting disinformation. Automation of AI and social media, introduced from 2025, should make everything more transparent and control misinformation. However, rapid technological change and the inherent inefficiencies of content-moderation mechanisms are important substantive obstacles. Thailand, for example, lies on the mid-range of tactics: vote counting, attention paid to demonstrations, and use of cameras. While there have been the issuance of a few prescriptive regulations, matters related to the ethical aspects of AI legislation are still being deliberated, waiting for the finalization of a regulatory statute. These measures have improved vote accuracy and monitoring efficiency, but ambiguities regarding policy and ethical concerns, especially regarding surveillance, call for further legislation. Although there are some limited applications of AI, such as electoral monitoring, its application is hindered by the country's weak regulatory regimes. The country's weak stance toward electoral transparency makes it vulnerable to authoritarian use of AI technologies for election manipulation and compromising democratic norms (Kingdom of Thailand, 2019). This paper represents the heterogeneous profile of AI in SEA's electoral governance landscape, and Indonesia's biometric vanguard, in particular, faces significant regulatory and ethical dilemmas. Insights derived from International IDEA (2025) and RSIS publications and from the realm of governance and electoral documents of the Asean offer a substantive basis for policymakers and stakeholders to demonstrate the new nexus between AI and dynamics of democracy and authoritarianism in the region.

Case studies

Several Southeast Asian countries have seen how AI makes a difference in their electoral management. The successful implementation of the most sophisticated and advanced biometric voter identification system in elections in Indonesia to tackle the electoral fraud has been able to ensure accurate voter registrations and eliminated number of bogus voters (Aminuddin, 2017), more than this, the biometric voter authentication system is AI powered that uses multiple modes of authentication like facial recognition, biometric scanning or Irish identification to find out the eligible voters (Lestari et al., 2025). AI was clearly visible in the 2019 Thai general elections. Thailand relied on AI to check voter IDs and predict how many people would vote during the election. Computer systems were used to study social media, find what voters are saying and guide political parties in directing their campaigns (Jalli et al., 2019). Instead, the election faced claims of cheating, and some said the use of AI systems likely encouraged favoritism for some political interests. From this example, we notice why it's crucial to ensure that AI is open and responsible to address incorrect use. Authorities are using AI in Myanmar to spot fraudulent actions in election processes (Croissant and Lorenz, 2018). For the 2020 general election, artificial intelligence was employed to spot different types of fraud, such as voting by people who should not be eligible (Iglesias et al., 2013). Elections were made easier with the help of AI, but the nation's authoritarian government and use of repression raised doubts about the purpose of AI in their elections. Because AI risks being used by the government to control elections in Myanmar, the country should strengthen its regulatory rules managing AI in elections (Subekti and Wahid, 2023). The country of the Philippines is also relevant because anyone can see how social media is often manipulated during its elections. Delivering targeted political ads and false news to people using algorithms has stood out as a prominent feature of recent Canadian elections (Reilly, 2014). In the 2016 presidential election, artificial intelligence helped distribute biased news and messages about candidates. Because artificial intelligence is used in social media campaigns to influence voting opinions in the Philippines, major ethical issues with AI in electoral matters and its effect on democracy must be raised (Shukri, 2023). AI has led to new chances and challenges when managing elections in Southeast Asia. AI tools are able to help make elections easier, more open and available to more people. Yet, as organizations work on these new technologies, challenges about influencing voters, using biased algorithms, and privacy arise (Park, 2011). Because AI is going to exert its influence in Southeast Asian elections, new rules are needed to ensure its technology can't threaten democracy's main principles (Suiter et al., 2018). The examples set by Thailand, Myanmar and the Philippines mean that elections across Southeast Asia are greatly influenced by AI, which is why we require innovative answers to profiting from AI while controlling its risks (Kassim, 2012).

Problems of voter manipulation and their link to ethics

Using Artificial Intelligence (AI) more in elections has produced doubts about how it might harm voters through repeating biases, providing false information and invading privacy. Because AI helps improve elections, not managing its use could be very damaging to democracy (Luzyanin, 2023). Below, we examine how biased algorithms, deceptive information and privacy problems might occur in elections with AI, providing examples of AI misuse for fraudulent voting in Southeast Asia.

Algorithmic bias

AI algorithms look at data in order to decide, so if the training data has bias, the outcomes can wind up making current inequalities even worse. Its influence concerns elections the most, as both representation and treatment should be even there. Systems that use artificial intelligence in electoral activities like voter registration, checking identities and campaigning often use information from data about demographics, voting behavior and social media (Pempel, 2013). Sometimes, just the nature of biases in the data used can trick algorithms into strengthening current inequalities. For example, software used in voter identification based on facial recognition seems to misidentify more women and people of color in places where most of the training data is made up of light-skinned images (Prokurat, 2014). As a result, there is a risk that marginalized communities can be either badly overlooked or confused, harming the fairness of electoral procedures. People worry that AI is adding to the problem of bias when it comes to political advertising (Kwon, 2010). Algorithms used in social media ads check many data and then show targeted ads to voters. By coincidence, political algorithms often reinforce pre-existing opinions, resulting in voters only ever listening to information that agrees with them, which contributes to more division (Satrio, 2019). As a result of AI, political campaigners are able to send different messages to people of different social or financial backgrounds, depending on where they live or what they interact with online. As a result, there is a danger that elections will produce more social division and reduce the chance for informed discussions and the exchange of ideas (SarDesai, 2018).

Disinformation and fake news

AI is also rightly accused of being used to share false information and disinformation. Machines powered by AI have been used to produce, forward and amplify blatantly wrong or deceptive content, mostly on social media. Thanks to its speed in analyzing who to target, AI makes it easier for those seeking to spread misinformation to try to influence how people vote. Many times, bots and algorithms controlled by AI help distribute content with political purposes, affect opinion and disrupt the voting process (Nohlen et al., 2004). During political campaigns, AI used in social media chooses and directs personalized content at voters that may not be reliable. Making fake news louder using technology can greatly confuse politics and erode citizens' trust in their democracy (Ostwald, 2017). At the time of the 2016 election, disinformation in Southeast Asia was evident in the Philippines. Automated bots on social media pushed out fake news and intentionally biased articles, mainly in favor of Rodrigo Duterte. Using these bots, a collection of fake accounts circulated made-up stories, over-the-top claims and false information to try to shape the way people voted. AI-based platforms spread disinformation to many people, which strongly impacted voters' behavior and concerned people about using AI for politics (Keum and Campbell, 2018).

The use of AI in Thailand lets politicians bombard certain groups of voters with information that was not true or only half the story. In the 2019 Thai general elections, politicians and their supporters boosted political messages through social media and used AI to spread fake stories against rival parties. Thailand's story shows that using disinformation online makes it harder to tell true from untrue information during elections (Aspinall, 2005). AI is not only involved in spreading disinformation and fake news during campaigning. Many Southeast Asian countries have seen state actors use AI to influence how the public thinks and to hush down protests. The military in Myanmar has taken advantage of social media platforms and AI algorithms to focus on ethnic minorities, dispense false news and curb those who want to change the government during the 2020 general election (Harahap, 2021). Artificial intelligence is being used to control voters unfairly in elections, which threatens democracy and raises new moral questions for its users in both political and autocratic regimes.

Surveillance and privacy

The growing adoption of AI in elections has created serious questions about protecting people's privacy. AI is able to collect, check and apply tons of private data, usually without the need for individuals' explicit consent (Harahap, 2019). For this reason, some worry about people who can watch voters, invade their privacy and use voters' personal data in dishonest ways for politics. AI is largely used to observe voter tendencies on social media sites by studying the huge amount of data gathered there. With the help of the data, political message makers prepare content that relates to individuals' political views, who they are and what they do online. Still, this practice worries many people since, in places where privacy protection is low or absent, it can lead to security breaches (Mobrand, 2020). In the Philippines, analyzing data with AI to target voters with individual political ads is worrying many people about privacy. Facebook and other platforms widely used in Philippine elections have faced criticism for letting campaigns use private user data to try to sway voters. Tracking and analyzing how voters behave online, with low levels of transparency, raises important questions of ethics involving consent, privacy and the chance for misuse (2017 Dissertation List, 2018). AI is now being used in Indonesia to keep an eye on voter actions online and on social media. In places with fewer rules for protecting data, it is common for people not to realize how political campaigns or outsiders might use their information (Nohlen et al., 2001).

Cases where manipulation has happened

Several Southeast Asian countries have noticed that AI and digital tools are being misused to sway voters around elections. In the Philippines, Thailand and Myanmar, AI is being used for political advantage in ways that worry ethics and regulatory officials. As was pointed out before, AI bots operating on social media in the Philippines spread disinformation during the 2016 presidential election (Iannone, 2022). The bots sent out untrue stories, emphasized false political information and influenced voters with political messages. Infecting public opinion with false information that is pushed through AI can seriously affect electoral fairness because voters are fed falsehoods (Lemonik Arthur, 2020). At the time of the 2019 General Election in Thailand, both political groups and disinformation campaigns used AI to customize and broadcast their ads to voters. Automated systems on platforms were exploited to make false news, promote biased views and gag opposition. AI use in political contests in Thailand points out that digital platforms and AI bring ethical problems to democratic elections in nations with problems of instability and censorship of the press (Harding and Leyland, 2008). Myanmar shows one of the most alarming ways AI has been involved in elections. The military government used new AI technologies on social media to push its own views, block expressions of disagreement and fill online discussions with misinformation just before 2020. By using AI to spy on people and influence votes in Myanmar, dictators revealed how dodgy methods could easily gain such regimes more influence over voters. Since artificial intelligence can be misused for repressive control when there is no democracy, such uses must be addressed ethically (Burnell and Ware, 2017). Because AI is used in electoral management, concerns should be raised about possible threats to democracy in Southeast Asia. Algorithmic bias, the sharing of incorrect information and the breach of privacy rules all threaten fair elections and individuals' rights. AI and digital technologies have the potential for misuse, as cases from the Philippines, Thailand, and Myanmar demonstrate, which means rules and principles are important to guide their use in voting (Cheng, 2003). As AI becomes more important in politics and elections, there is a strong need for the triple action of policymakers, election officials and tech developers to keep AI use transparent, fair and ethical while honoring democracy and safeguarding voters' freedom.

Policy and regulatory frameworks

The use of AI in election management in Southeast Asia has spurred the demand for strong guidelines to control how AI is used, so that election processes are protected. Even though a few nations in Southeast Asia are working on AI rules for elections, the picture is mixed, since rules and enforcement change depending on where you are. We will review the current rules for AI in elections across Southeast Asia, see how they measure up to the best practices globally and make suggestions to offer better protections for risks linked to AI (Gartenstein-Ross, 2015). Elections and the use of AI are still in early stages in Southeast Asia, since several governments have not yet formed clear and similar guidelines. There are clear cases where countries are introducing or are about to implement rules around how digital technologies and AI play a role in elections. Among its measures, the Indonesian government has put rules on digital services and the use of AI-based voter registration systems in its elections (Zhou et al., 2014). Under the guidance of the General Election Commission of Indonesia (KPU), all these systems follow both the Data Protection Law of Indonesia and the country's election rules. However, laws about AI's broader impact on elections have not been fully built out, as people are concerned about the security of data, clarity of algorithms and AI's ability to affect how people vote. Even though the government is working to secure elections with new identification systems and digital voting, unclear rules about AI's ethics in these areas create big holes in regulations (Ginsburg, 2008). AI is being used in Philippine elections mainly under the direction of the Commission on Elections (COMELEC) in electronic voting and voter registration. AI technologies have been applied in biometric enrolment in the country to make sure voter lists are both secure and reliable (Bangun, 2016). Guiding standards in data privacy from the National Privacy Commission (NPC) exist, yet there is not much law governing the ethical use of AI for elections. There are growing fears that ads using AI to target political messages in elections can trick voters, but few laws prevent such practices (Kiyohara et al., 2018).

In Thailand, the Election Commission of Thailand (ECT) uses digital solutions such as AI to both track and predict the actions of voters online and enforce safeguards on reliable information. Undeniably, there are no complete laws in the country to manage how AI is used in political programmes. The law covers some basic digital security questions, yet it leaves out the ethics of AI in elections, covering things like data privacy, misleading information and possible bias in AI (Kendall-Taylor and Frantz, 2016). On top of that, since the rules for political ads on social media are unclear, AI-based methods of influencing people's opinions are more likely to succeed (Cheng and Chu, 2017). Even though many countries in Southeast Asia are beginning to manage AI in elections, the legal systems are not strong enough to address all challenges associated with AI in electoral management. Because every country has its own rules, regions in Southeast Asia do not have standard policies, which exposes many countries to risks associated with AI (Anas, 2022).

Best practices from around the world

Not like Southeast Asia, the European Union and the United States have made more complete rules for AI in elections, and their approaches can teach Southeast Asia a lot. Because of the European Union's GDPR, personal data within AI-based elections is protected, which prevents both violations of voter privacy and misuse of information. Although the GDPR was not written with elections in mind, its strong data rules cover all sectors, including those responsible for holding elections. Such systems are obliged by the law to ensure transparency, respond to accountability and avoid discrimination, as personal data processing is done legally and without risk. Besides, the European Commission's Digital Services Act (DSA) and Digital Markets Act (DMA) demand that all aspects of digital platforms be open to users, that content is moderated correctly and that algorithms are made responsible (UNHCR, 2000). They make high demands when it comes to AI ethics in politics and offer a good model for Southeast Asia to follow in its work to create AI regulations for elections (European Commission, 2020). In the United States, the Federal Election Commission handles election rules covering how political campaigns can use digital channels. Though the U.S. lacks broad AI rules for elections, the BCRA and its revisions made rules for political advertising on the internet. Thanks to the regulators, political groups now have to share information about who is helping them financially and how their ads are focused. Still, there are a few rules to regulate the use of targeted ads that reach single voters at the micro-level. Concerns about how AI influences elections have inspired certain U.S. states to set new rules to make digital advertising more open and to regulate ethical concerns of AI in political campaigns (Federal Election Commission, 2021). Experts from around the world say that AI should be used transparently, fairly and with accountability during campaigns and elections (Hsiao, 2010). Thanks to these rules, the use of these systems is under oversight, helping to cut down on risks of exploitation, bias or mishandling data. Southeast Asian nations have made forward movement, yet there is still a lack of laws and action to fight against cybercrime.

Southeast Asian countries ought to build new, detailed regulations for AI use in elections. Part of this effort is to make new laws about AI's acceptable use in voter registration, political marketing and monitoring elections (Zenn, 2013). Every framework for voting should care about voter privacy, data security, and avoid allowing AI to influence the outcome of votes. Indonesia, the Philippines and Thailand could gain from putting in place AI-specific rules that follow the example of best international standards in transparency, accountability and non-discrimination. To sharpen voter privacy, the lawmakers in Southeast Asia should update their statutes based on the GDPR rules of the EU. They must be designed to meet tough privacy requirements and give power to voters over their own personal information. There ought to be detailed rules for gathering, processing and storing voters' information and clear punishments when these rules are broken (Taxamo, 2021). Require rules that make political campaigns in Southeast Asia explain how electronic campaigns through AI are used to target voters and place ads. Much like the EU and the U.S., all Southeast Asian countries should make transparency a rule for AI-powered political ads. For example, we should require ads to list the source, specify how people are selected to see them and how their data is analyzed. Governments and international organizations should partner to teach the public about artificial intelligence and the risks linked to using it in elections. Helping people understand how AI impacts voting, tricks people online and worries about privacy is necessary to preserve fair elections now (Fukuoka, 2013). Innovative voting systems should have an overseeing body that keeps AI use in the election or campaign process transparent and under control. It is the job of this body to oversee compliance with law, investigate allegations of voter manipulation and present policy improvements. Regional cooperation is needed (Mair, 1960). Southeast Asian countries should decide on standard guidelines for AI in elections by using ASEAN and related groups. Sharing knowledge, aligning rules and setting a common regional approach to AI will help Southeast Asia overcome the challenges AI brings to electoral processes around the region (Perron, 2010). There are many new risks and opportunities connected to AI systems in Southeast Asia's elections. Improving election procedures is the goal of AI, but any voter manipulation, biased algorithms or breaches of privacy can be big problems caused by AI. AI should help Southeast Asian elections in a good and fair way through the drafting of strong rules, more secure data protection rules, explicit campaign content and shared knowledge of its use. When Southeast Asian countries collaborate globally and with their neighbors, they can pass rules to keep democracy and guide how everyone makes use of AI in elections (Anglin, 1961).

Legal and regulatory frameworks and international best practices of AI in electoral management

Strong legal and regulatory frameworks are necessary to balance the innovation of AI in the electoral administration process with the protection of the democratic integrity within it. Within Southeast Asian countries, there are both glimpses of progress and significant shortcomings in this regulatory spectrum, illustrating the potential and complexities of utilizing AI in the electoral sphere. In the Philippines, the Data Privacy Act of 2012 (RA No. 10173) and its implementing Guidelines issued by the National Privacy Commission provide stringent personal data protection measures with specific consent requirements, principle of data minimization and notification of data breaches. This fact is critical to implementing the Commission on Elections Automated Fingerprint Identification System (AFIS), which was to have transparency and accountability measures as mandatory provisions (Republic of the Philippines, 2012). Indonesia's recent but extensive Personal Data Protection (PDP) Law (2022) and Government Regulation Number 71/2019, which state mandatory requirements for data controllers, the regulation of cross-border transfers and harsh penalties, directly tackle the existing flaws in the government's eKTP biometric database (Republic of Indonesia, 2022; Cabinet Secretariat of the Republic of Indonesia, 2019).

In Thailand's regulatory regime, the Personal Data Protection Act (PDPA) in 2019 brought data protection compliance in line with international standards and prescribed Data Protection Impact Assessments (DPIA) for AI-based election technologies while empowering the data subject and sanctioning for non-compliance. In line with this, the Election Commission's Rules on Electronic Campaign (Komisi Pemilihan Umum, 2019) require accountability and transparency of computerized political advertising with disclosure of algorithmic targeting mechanisms. Myanmar: The emerging political and legal landscape has no comprehensive DPA; however, the existing Telecommunications Law of 2013 emphasizes the protection of subscriber confidentiality, and pending cybersecurity legislation offers an incipient statutory framework relevant to emerging AI electoral technology (Republic of the Union of Myanmar, 2013).

There are many practical examples around the world for AI in electoral governance which set instructive precedents. An excellent example of transparency in this regard is the e-voting system in Estonia, which has free open-source codebases with frequent independent audits and ensures algorithm transparency to help build public trust (Vassil, 2020). The UK's Election Commission has made very specific provisions forbidding digital political campaigning, including the sharing of algorithmic approaches to give information about targeted ad techniques, as well as post-campaign reporting, strengthening electoral transparency (Feldman, 2020). Conversely, in the United States Voluntary Principles on Election Modernization, it is mandated that robust pre-deployment audits of algorithms and ongoing monitoring of the product in real-time are required to quickly find and rectify abnormalities (Binns, 2020). By institutionalizing these mechanisms, the paper argues for the creation of an Independent AI Electoral Oversight Agency in the region, which would be given statutory authority to establish transparency oversight rules, ongoing audits, and compliance with the norms with respect to data protection standards and existing laws. In light of the above, this paper advocates the creation of this agency to operationalise accountability, promote resilience against AI bias and the misunderstanding or misuse of data, and manage the ethical governance of AI in electoral processes. Such an establishment would work as a construction of democratic integrity in the virtual era.

Discussion

Implications of the main findings

The study examines what changes Artificial Intelligence (AI) brings to the management of elections in Southeast Asia, looking at both pros and cons. According to the experts, there are many benefits to AI alongside plenty of ethical concerns. Through AI, the process of voter registration has been improved, and confirming voter identities and watching voting has been made easier at a national level. Since there is AI, elections in both Indonesia and the Philippines are much better protected from fraud and incorrect votes. AI can also play a role in checking elections, since tools that watch over the process spot mistakes and help predict how voters are expected to vote, which can keep the process clear and compelling. What is more, thanks to AI, political campaigns now have a greater reach, as personalized ads and information can be delivered to individual voters, so the hope is that more people become involved in the process. At the same time, bringing AI into how elections are run brings several serious risks. Many leaders are worried about algorithmic bias because if a system is not controlled correctly, it could make inequality worse for underserved people. Mainly because of targeting by AI, there is a threat to the quality of elections caused by increasing political division. Besides, AI used in social media helps misguide voters by making fake news and disinformation reach many people. It is also clear from the findings that, while some Southeast Asian nations have begun to govern AI in elections, the regulations in place are either not adequate or not standardized. Since there are no clear guidelines on how to conduct elections if they study policies on AI in their electoral management, the region is at risk of their misuse. The rules currently in place do not handle all the problems AI brings, which is why better regulatory guidance is needed to prevent abuses of AI in elections.

Implications for electoral integrity

AI is expected to have a big impact on the integrity of elections in Southeast Asia. For one, AI can lift the standards for the efficiency, safety and transparency within elections. Thanks to AI, people can find it simpler to enroll in elections, poll results can be checked more closely, and election outcomes can be estimated more accurately. In addition, relying on AI for early detection of problems and fraud based on biometric methods may raise public trust in elections. Current AI risks, for example, voter manipulation, biased algorithms and invasion of privacy, could actually hurt the election system. The chance to use AI in campaigns, targeted ads and profiling voters leads to significant questions about the fairness of choosing our politicians. Since these risks mix with insufficient regulation in Southeast Asia, far too many people may turn away from seeing democracy as legitimate, which could result in fewer supporters voting. Electoral integrity is urgently questioned in authoritarian countries, because AI might help extend, rather than reveal, power. In Myanmar, people ask whether AI is being used to help governments become more authoritarian. With proper measures, anybody hoping to control elections or restrict freedom of opinion could make the democratic foundations of elections weaker.

Areas for future research

Although this research gives us a lot of details about how AI supports election administration in Southeast Asia, some important topics still leave room for further exploration. First, finding the best ways to regulate AI in election processes is a topic that needs much more study. Further analysis is needed in terms of legal and ethical aspects of AI in elections, including ways countries can develop policies to promote election fairness, transparency, and avoid the abuse of AI. Studies need to look at which strategies help elections avoid the downsides related to AI being used without total fairness. The way digital platforms and AI influence vote behavior and democracy over the term needs more study. Since AI is now involved in many social media political campaigns, research should focus on how political advertising and disinformation impact the choices of voters, public opinion, and democracy. Democracy and election integrity research shows promise by observing changes in affected countries over a long period. We need to explore more the link between AI and digital literacy. AI's role during elections is not well understood by many in Southeast Asia. Studies should see how well people in slower-adoption nations can use AI in the voting process. They ought to study its effectiveness to let people know, to teach them technology and to prepare them for the risks AI can have in elections. We need to think seriously about using AI in surveillance and for tracking votes. With AI being more commonly used to keep track of voting, researchers must explore what ethical problems arise from this practice. Finding out how AI can be used to guarantee both voter privacy and fair voting is a priority for many researchers. AI might be helpful for election management in Southeast Asia, despite the risks it creates that should be looked at. Although AI makes elections more open, reduces fraud and streamlines administrative tasks, the difficulties resulting from voter manipulation, biased algorithms, disinformation and privacy are still a big issue for election management. AI ought to strengthen democracy in Southeast Asia, so the area has to boost its rules, ensure elections are transparent for all and help people better understand both the benefits and risks of AI used in elections. Dealing with these challenges allows Southeast Asia to benefit from artificial intelligence and support the central values of democracy in elections. It is important to keep good, open practices in elections by exploring AI's influence on election rules and how people see AI.

Conclusion: safeguarding democracy in the age of AI

The current research paper claims that Artificial Intelligence in Southeast Asia has significant potential to increase the efficiency, transparency and accessibility of the electoral management processes, but, at the same time, its use is also negatively affected by challenges that threaten electoral integrity. The examples of the Philippines, Myanmar, and Thailand show that without strict regulatory policies, transparent algorithms, and close monitoring, AI has the potential to exacerbate threats of inequality on a large scale due to biases, to manipulate and to utilize personal data. Registering voters more quickly, bringing more transparency to vote counting and delivering political ads better are all possible with AI. Because of these improvements, elections across Latin America will be simpler, safer and more accurate, making democracy stronger everywhere. Besides, depending on AI for strategy, political parties can influence how voters feel and decide, and this approach usually isolates people and makes choosing harder. When artificial intelligence tools are used to threaten elections, disinformation and privacy become even bigger problems. That is the reason why the use of AI in elections should be subject to clear ethical rules and be completely open. It's important for countries in Southeast Asia to design clear and complete policies that address problems related to AI, especially concerning fairness in algorithms, data safety during elections and controlling digital politics. Such frameworks ought to highlight being transparent and answerable, so there are tools to monitor and manage AI solutions used for elections. In addition, society, politicians and policymakers all require more information concerning the ethical implications of AI when it is used in elections. To do this, we should also work on making and putting into place policies that protect against the misuse of AI while making sure AI is used to improve voting systems. Efforts by public authorities, election commissions and international organizations ought to concentrate on creating just one safe use of AI in elections. Basically, AI could solve many electoral challenges in Southeast Asia, though we must be careful about the specific problems it faces.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

SS: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. NG: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. AY: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing.

Funding

The author(s) declare that no financial support was received for the research and/or publication of this article.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Abbreviations

AFIS, Automated Fingerprint Identification System; AI, Artificial Intelligence; BCRA, Bipartisan Campaign Reform Act; COMELEC, Commission on Elections (Philippines); DMA, Digital Markets Act; DPIA, Data Protection Impact Assessment; DSA, Digital Services Act; ECT, Election Commission of Thailand; EU, European Union; FEC, Federal Election Commission (United States); GDPR, General Data Protection Regulation; KPU, Komisi Pemilihan Umum (General Election Commission of Indonesia); NPC, National Privacy Commission (Philippines); NGO, Non-Governmental Organization; PDPA, Personal Data Protection Act (Thailand); PDP Law, Personal Data Protection Law (Indonesia).

References

2017 Dissertation List (2018). PS Polit. Sci. Polit. 51, 713–725. doi: 10.1017/s1049096518000823