16 February 2024
CO24023 | AI and Elections: Lessons for Southeast Asia
SYNOPSIS
2024 will see consequential general, legislature, and presidential elections in several countries around the world. In Southeast Asia, the Indonesian elections on 14 February will be particularly significant given the country’s size and economic heft. Unprecedented advancements in artificial intelligence, however, have led to an anticipated surge in election-related disinformation. To address this challenge, Southeast Asian countries must adopt a comprehensive strategy to safeguard election integrity against evolving digital threats.
COMMENTARY
2024 has been dubbed the ultimate election year. In January, Taiwan and Bangladesh went to the polls and decided on the winners of the seat of power. In February, there is the massive Indonesian general election and the first Cambodian Senate election following the retirement of long-serving Prime Minister Hun Sen.
As artificial intelligence (AI) has made unprecedented advancements in recent years, it can shape the outcomes of these elections and pose a threat to democracy. Experts foresee a surge in election-related disinformation on a global scale.
Impact of Disinformation on US Elections
Although there is much attention on the 2024 elections in Russia (March), India (April), and the European Parliament (June), the most significant focus and impact of AI-generated disinformation is arguably on the US elections in November.
During the US presidential election in 2016, computational propaganda and bots distributed manipulated content widely on social media, shaping narratives and hardening viewpoints. Hyper-targeted ads based on stolen data allowed the campaign of presidential candidate Donald Trump to micro-target and suppress opposition turnout. Disinformation and propaganda tactics have demonstrably shifted voter attitudes and turnout, influencing final margins.
During the New Hampshire primaries held last month, an AI-powered robocall impersonating US President Joe Biden was circulated discouraging Americans from voting while urging them to save their votes for November instead. Although the full impact of this incident is yet to be seen, the potential risks are substantial as the US already has a “primary problem”, with only a small fraction of eligible voters turning up to decide which candidates appear on the general election ballot.
While it is easy to blame the less educated and less affluent members of the electorate for falling for disinformation tactics, it is important to recognise that perpetuating the “dumb voter” trope not only marginalises specific social classes but also plays into the hands of populist leaders. Furthermore, deepfakes are not a class-based problem, as even CEOs of large companies fall prey to simple voice-generating AI deepfake scams.
The use of deceptive deepfake technology to spread falsehoods during a crucial democratic process offers a glimpse into how AI could be strategically weaponised to sow confusion and influence election outcomes if not checked. The Biden robocall incident is a wake-up call for democracies worldwide to enact safeguards and equip citizens with more robust media literacy skills.
Southeast Asian countries could benefit from monitoring events leading up to the US presidential election in preparation for similar attacks on their elections.
Deepfakes and the Indonesian General Election
The Indonesian election held on 14 February could provide some useful lessons. One surprising aspect of the election was the hands-off stance that the Indonesian General Elections Commission (Komisi Pemilihan Umum in Indonesian, abbreviated as KPU) took regarding the use of AI in campaigning, leaving voters to their own devices when separating facts from faked information.
This should have been a concern as Indonesia had been plagued by AI-powered deepfakes ranging from fabricated health claims supposedly made by the former Minister of Health to a political deepfake video of the late Indonesian President Suharto urging voters to vote for Golkar candidates.
With the emergence of AI-powered propaganda, the 2024 Indonesian election presented an opportunity to observe and analyse the use of AI in disseminating political propaganda and disinformation.
Insights gained on the AI technologies employed, the nature of misinformation propagated, the effects of the KPU’s non-interventionist stance, and the effectiveness of countermeasures such as the new AI Ethics Circular announced by Indonesia’s Ministry of Communication and Information (KOMINFO), will be useful for other Southeast Asian countries. The lessons learnt could be helpful in the formulation of strategies to safeguard their respective electoral processes.
Lessons for Southeast Asia
Insights from the US and Indonesian elections can help inform Southeast Asia as it gears up for a series of polls. The use of AI in orchestrated disinformation campaigns is a worrisome trend and demands urgent attention, particularly in Southeast Asia where fake news is prevalent.
This challenge is amplified by high social media penetration in the region, making it a fertile ground for the spread of disinformation. In this new landscape, Southeast Asian nations must prioritise the formulation of comprehensive strategies to counter AI-driven disinformation. This requires the following steps:
(i) The deployment of advanced technological solutions equipped with local information sources and using AI and data-driven approaches that can detect and neutralise AI-generated disinformation. Current AI fact-checking measures, however, are not foolproof, particularly for context-dependent content.
(ii) A concerted effort to enhance public literacy in AI. Falsehoods that confirm biases can be mitigated by bolstering public education on how knowledge is produced. Information professionals can play a key role in this by harnessing the multiplier effect. Training the general population in public spaces such as libraries can produce ripple effects leading to a more knowledgeable electorate. Practical solutions to AI-powered deepfakes in the context of Southeast Asia have been discussed at a workshop held in Kuala Lumpur in 2020, empowering media practitioners and journalists to step up and share tools, skills, and practices in combatting AI-powered disinformation.
(iii) Drawing lessons from past experiences. While disinformation is just one aspect of deeply-rooted problems in Philippine politics, the 2022 election won by Ferdinand Marcos, Jr., saw the use of similar disinformation tactics on social media employed by his predecessor Rodrigo Duterte during the 2016 elections. This could have been prevented.
(iv) Southeast Asian nations should ensure that their AI policies are aligned with the interest of the electorate, have room for electoral reforms, and are designed to protect the integrity of elections. A multi-stakeholder approach involving academics, industry players, the media, and the public alongside policymakers can provide a well-rounded policy for the benefit of the majority. An example of this is to be seen in Singapore’s initiative to invite international feedback on its proposed Model AI Governance Framework for Generative AI. Outcomes of the Singapore approach could offer insights for other countries drafting their own AI policies.
Conclusion
In navigating complex election-related challenges brought about by AI, Southeast Asian countries can benefit from a holistic approach that combines technology, robust AI literacy skills, willingness to learn from past events, critical observation of global and regional electoral trends, and responsible policy formation. This integrated strategy is important to ensure that elections in Southeast Asian countries are free and fair and strengthened against evolving digital threats.
About the Authors
Dr Karryl Sagun-Trajano is a Research Fellow for Future Issues and Technology (FIT) at the S. Rajaratnam School of International Studies (RSIS), Nanyang Technological University (NTU), Singapore. Dr Nuurrianti Jalli is an Assistant Professor of Strategic Communications at Oklahoma State University and a Visiting Research Fellow at ISEAS-Yusof Ishak Institute, Singapore.
SYNOPSIS
2024 will see consequential general, legislature, and presidential elections in several countries around the world. In Southeast Asia, the Indonesian elections on 14 February will be particularly significant given the country’s size and economic heft. Unprecedented advancements in artificial intelligence, however, have led to an anticipated surge in election-related disinformation. To address this challenge, Southeast Asian countries must adopt a comprehensive strategy to safeguard election integrity against evolving digital threats.
COMMENTARY
2024 has been dubbed the ultimate election year. In January, Taiwan and Bangladesh went to the polls and decided on the winners of the seat of power. In February, there is the massive Indonesian general election and the first Cambodian Senate election following the retirement of long-serving Prime Minister Hun Sen.
As artificial intelligence (AI) has made unprecedented advancements in recent years, it can shape the outcomes of these elections and pose a threat to democracy. Experts foresee a surge in election-related disinformation on a global scale.
Impact of Disinformation on US Elections
Although there is much attention on the 2024 elections in Russia (March), India (April), and the European Parliament (June), the most significant focus and impact of AI-generated disinformation is arguably on the US elections in November.
During the US presidential election in 2016, computational propaganda and bots distributed manipulated content widely on social media, shaping narratives and hardening viewpoints. Hyper-targeted ads based on stolen data allowed the campaign of presidential candidate Donald Trump to micro-target and suppress opposition turnout. Disinformation and propaganda tactics have demonstrably shifted voter attitudes and turnout, influencing final margins.
During the New Hampshire primaries held last month, an AI-powered robocall impersonating US President Joe Biden was circulated discouraging Americans from voting while urging them to save their votes for November instead. Although the full impact of this incident is yet to be seen, the potential risks are substantial as the US already has a “primary problem”, with only a small fraction of eligible voters turning up to decide which candidates appear on the general election ballot.
While it is easy to blame the less educated and less affluent members of the electorate for falling for disinformation tactics, it is important to recognise that perpetuating the “dumb voter” trope not only marginalises specific social classes but also plays into the hands of populist leaders. Furthermore, deepfakes are not a class-based problem, as even CEOs of large companies fall prey to simple voice-generating AI deepfake scams.
The use of deceptive deepfake technology to spread falsehoods during a crucial democratic process offers a glimpse into how AI could be strategically weaponised to sow confusion and influence election outcomes if not checked. The Biden robocall incident is a wake-up call for democracies worldwide to enact safeguards and equip citizens with more robust media literacy skills.
Southeast Asian countries could benefit from monitoring events leading up to the US presidential election in preparation for similar attacks on their elections.
Deepfakes and the Indonesian General Election
The Indonesian election held on 14 February could provide some useful lessons. One surprising aspect of the election was the hands-off stance that the Indonesian General Elections Commission (Komisi Pemilihan Umum in Indonesian, abbreviated as KPU) took regarding the use of AI in campaigning, leaving voters to their own devices when separating facts from faked information.
This should have been a concern as Indonesia had been plagued by AI-powered deepfakes ranging from fabricated health claims supposedly made by the former Minister of Health to a political deepfake video of the late Indonesian President Suharto urging voters to vote for Golkar candidates.
With the emergence of AI-powered propaganda, the 2024 Indonesian election presented an opportunity to observe and analyse the use of AI in disseminating political propaganda and disinformation.
Insights gained on the AI technologies employed, the nature of misinformation propagated, the effects of the KPU’s non-interventionist stance, and the effectiveness of countermeasures such as the new AI Ethics Circular announced by Indonesia’s Ministry of Communication and Information (KOMINFO), will be useful for other Southeast Asian countries. The lessons learnt could be helpful in the formulation of strategies to safeguard their respective electoral processes.
Lessons for Southeast Asia
Insights from the US and Indonesian elections can help inform Southeast Asia as it gears up for a series of polls. The use of AI in orchestrated disinformation campaigns is a worrisome trend and demands urgent attention, particularly in Southeast Asia where fake news is prevalent.
This challenge is amplified by high social media penetration in the region, making it a fertile ground for the spread of disinformation. In this new landscape, Southeast Asian nations must prioritise the formulation of comprehensive strategies to counter AI-driven disinformation. This requires the following steps:
(i) The deployment of advanced technological solutions equipped with local information sources and using AI and data-driven approaches that can detect and neutralise AI-generated disinformation. Current AI fact-checking measures, however, are not foolproof, particularly for context-dependent content.
(ii) A concerted effort to enhance public literacy in AI. Falsehoods that confirm biases can be mitigated by bolstering public education on how knowledge is produced. Information professionals can play a key role in this by harnessing the multiplier effect. Training the general population in public spaces such as libraries can produce ripple effects leading to a more knowledgeable electorate. Practical solutions to AI-powered deepfakes in the context of Southeast Asia have been discussed at a workshop held in Kuala Lumpur in 2020, empowering media practitioners and journalists to step up and share tools, skills, and practices in combatting AI-powered disinformation.
(iii) Drawing lessons from past experiences. While disinformation is just one aspect of deeply-rooted problems in Philippine politics, the 2022 election won by Ferdinand Marcos, Jr., saw the use of similar disinformation tactics on social media employed by his predecessor Rodrigo Duterte during the 2016 elections. This could have been prevented.
(iv) Southeast Asian nations should ensure that their AI policies are aligned with the interest of the electorate, have room for electoral reforms, and are designed to protect the integrity of elections. A multi-stakeholder approach involving academics, industry players, the media, and the public alongside policymakers can provide a well-rounded policy for the benefit of the majority. An example of this is to be seen in Singapore’s initiative to invite international feedback on its proposed Model AI Governance Framework for Generative AI. Outcomes of the Singapore approach could offer insights for other countries drafting their own AI policies.
Conclusion
In navigating complex election-related challenges brought about by AI, Southeast Asian countries can benefit from a holistic approach that combines technology, robust AI literacy skills, willingness to learn from past events, critical observation of global and regional electoral trends, and responsible policy formation. This integrated strategy is important to ensure that elections in Southeast Asian countries are free and fair and strengthened against evolving digital threats.
About the Authors
Dr Karryl Sagun-Trajano is a Research Fellow for Future Issues and Technology (FIT) at the S. Rajaratnam School of International Studies (RSIS), Nanyang Technological University (NTU), Singapore. Dr Nuurrianti Jalli is an Assistant Professor of Strategic Communications at Oklahoma State University and a Visiting Research Fellow at ISEAS-Yusof Ishak Institute, Singapore.
