The Dark Side of Generative AI in Prabowo Subianto’s Presidential Campaign
In the months leading up to Indonesia’s general election in February 2024, AI-generated cartoons of presidential candidate Prabowo Subianto covered the country’s streets and social media platforms. The politician was portrayed as a doe-eyed grandfather who danced with his cat Bobby, quickly becoming a beloved online figure among supporters who branded him as “gemoy,” a term for anything cuddly and adorable.
Despite the guise of an avatar, Prabowo’s past looks significantly different from this positive characterization. As the current defense minister and former military general, he is accused of committing human rights abuses during former Indonesian President Suharto’s thirty-year dictatorship, leading to his subsequent exile and a generally feared image among the public. Now, after an online makeover, Prabowo will be sworn into office in October 2024 as Indonesia’s president-elect.
Prabowo’s campaigning was set against the backdrop of Indonesia’s young, social media-savvy voters. Prabowo held more than 60% of the Gen Z vote and 42% of the millennial vote, indicating his appeal among Indonesia’s younger cohorts. Thus, many of his newer followers are often uninformed about his policies and political background because they were not alive during the time where he was in power. This allows them to be easily influenced by his social media gimmicks. The ways in which Prabowo’s previous life was buried from his current public image offer a glimpse of how generative AI may transform large-scale political campaigning. The lack of governmental oversight regarding the use of AI in political campaigns has enabled politicians to weaponize generative AI to create historically revisionist content during election seasons—a trend that poses a serious threat to Indonesia’s democracy and public discourse.
Prabowo’s past political career contrastingly exemplifies his authoritarian leanings, particularly during his tenure as a military general from 1974 to 1998. At the height of his power, pro-democracy activists faced egregious violations and mysterious disappearances, with 22 reported abductions of activists and 13 cases of enforced disappearances. Although clear evidence suggests that Prabowo participated in and instigated these crimes, he has never faced legal repercussions in Indonesia, allowing his actions to be met with legal impunity by the Indonesian public.
Capitalizing on the fact that most of Indonesia’s Gen Z youth, who comprise 27.94% of the population, were born before these crimes were committed, Prabowo has actively tried, in the words of Professor Zachary Abuza “to whitewash his egregious past and reinvent himself with a softer, cuddlier image.” As such, his social media team focused on campaigning through TikTok and Facebook—both apps popular among his younger demographic. Prabowo’s avatar was made by Midjourney Inc., an American company used by political candidates to build interactive bots, generate AI portraits, and target voters. Prabowo’s campaign team made extensive use of the social media platforms’ services, spending over $144,000 on advertising in the three months leading up to the election.
Despite the successes of Prabowo’s campaign, AI-generated content has also ruined the political careers of other Indonesian candidates. Basuki Tjahaja Purnama, known as “Ahok,” was the governor of Jakarta in 2014. Ahok was strongly supported by digitally savvy groups, initially viewing social media as an asset for his reformist agenda. However, the tides turned when a manipulated video of Ahok’s speech referencing the Qur’an sparked public outrage, leading to protests and the promotion of anti-Chinese and anti-communist sentiment online. The public blowback from Ahok’s manipulated video resulted in a sectarian division within the 2017 Jakarta elections, in which voters were forced to align themselves with candidates of the same religious background. Ahok’s controversy highlights the volatility of generative AI, demonstrating how an asset can so easily become a weapon.
In the midst of Prabowo’s campaign, an AI generated deep fake video of late Indonesian dictator, Suharto emerged. In the video, Suharto’s voice and face are cloned, and he talks about the importance of voting for a member from the Golkar political party – the same party Prabowo is the frontrunner for. Just weeks before the election, Golkar brought a long-deceased leader back into the public eye, making its intentions clear: to align voters with the party still deeply tied to Suharto’s legacy. Golda Benjamin, a campaign manager at a US digital rights non profit expressed how “Deepfakes can greatly influence an election, especially its results.” This is because a deep fake can “easily reach millions in seconds, swaying and manipulating (millions of) voters.”
As Prabowo and Ahok’s campaign demonstrates, there is a crucial need to regulate the use of generative AI in electoral politics. Indonesia’s current legal frameworks are insufficient in promoting an environment of transparency and accountability. There are no specific regulations mandating the disclosure of expenditures for social media campaigns and marketing teams are not subject to governmental oversight. The lack of regulation on AI usage in political campaigning enables politicians like Prabowo to systemically deceive voters, hiding behind innocent cartoons to promulgate their totalitarian views and disregard for human rights.
A comprehensive strategy for tackling this problem should equally focus on education and digital literacy. Nadia Naffi, a professor in educational technology, states that “the role of education is indispensable in the fight against deep fakes.” As governments struggle to combat new forms of digital disinformation, educating the population on how to navigate disinformation and new technologies, guided through civics and media literacy programs, seems to be the most sustainable solution.
To further mitigate the abuse of generative AI, Indonesia’s broader regional bloc, the Association of Southeast Asian Nations (ASEAN), should establish a shared understanding of the misuse of AI in politics. ASEAN has created a guideline for AI governance and ethics, but creating a more rigid regulatory framework on ethics is difficult when applying it to state-specific situations. One weakness of these guidelines, experts say, is that “framing AI discussions through a predominantly business-focused lens diminishes the wider context in which AI operates” and thus cannot provide meaningful guarantees of social equity. As a whole, ASEAN’s regulatory environment focuses on “non-interference” which sets regional norms that emphasize member states’ sovereignty and self-governance. However, such permissive enforcement leaves room for potential misuse of AI, necessitating ASEAN to take on a more active role in policing its usage. A danger in this relatively laissez-faire policy, is that deep fakes of current political leaders can circulate, potentially threatening political alliances between countries.
Social media in Southeast Asia has shown itself to be a platform that continually forgives, offering second chances to political candidates who have committed human rights abuses in the past. Prabowo’s election has ushered in a new era of Southeast Asian politics, which weaponizes social media to craft revisionist narratives and cause online discourse rampant with disinformation. While we have been distracted by dancing cartoon politicians, social media is drastically altering our understanding of politics–and the leaders meant to represent us.
Tara Tulshyan (CC ’27) is a staff writer at CPR studying economics and political science.