In this webinar, Ms Khoirunnisa Nur Agustyati and Dr Ismail Fahmi explored the use of Artificial Intelligence (AI) to generate content and strategies in electoral campaigns and the possible solutions to address the spread of disinformation on social media channels.
INDONESIA STUDIES PROGRAMME WEBINAR
Friday, 21 July 2023 – Artificial Intelligence (AI) has emerged as a tool to influence political campaigns through the creation of political messages in response to campaign developments. Ms Khoirunnisa Nur Agustyati (Perludem) and Dr Ismail Fahmi (Indonesian Islamic University), two esteemed experts on the electoral process and AI, shared their observations on information disorder in the polarized Indonesian society and how AI could transform the political campaign in the upcoming 2024 elections. Dr Siwage Dharma Negara, Senior Fellow and Co-Coordinator of the Indonesia Studies Programme at ISEAS, moderated this session.
Ms Khoirunnisa examined the role of civil society in tackling information disorder among Indonesians during election periods. Reflecting on the 2014 and 2019 elections in the country, she highlighted how rising disinformation, spread by the use of social media platforms, had led to increased polarization among voters in Indonesia. She cited the differences in social media usage in the two previous General Elections. For the 2014 elections, social media was used primarily to shape people’s opinions toward the elections, with hoaxes only appearing on election day itself. This was different in the 2019 elections where hoaxes circulated widely throughout the election process and were used to sustain polarizing opinions toward certain political parties. The rampant spread of misinformation had also led to distrust toward the electoral system and the legislative body, evidently seen in the riots that occurred after the announcement of the election results in 2019. Ms Khoirunnisa, therefore, expressed concerns that the lack of regulation to tackle disinformation could lead to massive mistrust toward the upcoming 2024 elections. Coupled with the introduction of AI, people in Indonesia could be overwhelmed with information, resulting in a higher risk of widespread misinformation.
Ms Khoirunnisa believed there is a need to increase transparency in social media usage for election campaigns. Besides, she also proposed the establishment of a digital democracy ecosystem that can detect, analyze and reveal disinformation. This includes engaging the help of civil society organizations as well as synergizing with other stakeholders (such as government and social media platforms) to prevent disinformation. Lastly, there is a need to strengthen voters’ resilience towards disinformation through education. She cited the 2019 election as an example where the debunking of hoaxes by the authority was proven ineffective once they were released on social media platforms. She, therefore, proposed the idea of conducting pre-debunking of misinformation before the 2024 elections. This would aid in educating the population to be mindful of misinformation, minimizing the risk of information disorder in the upcoming elections.
Going deeper into tactical strategy, Dr Ismail Fahmi explored the use of artificial intelligence in political campaigning and its impact on society. He began by providing a brief introduction on the development of AI approaches, emphasizing that the learning process of any AI approach relied heavily on the input data. This in turn could lead to misuse, especially when data being inputted into the AI system are often biased in nature. People could use these biased contents to drive certain perspectives and sentiments towards the presidential candidates. That being said, presidential candidates could also use AI approaches in a positive manner, mainly to generate content for their campaigns. This would aid in increasing the frequency of their political messages on these social media platforms, allowing them to spread to a broader audience. This also helped in boosting the visibility of these political candidates on social media platforms.
Similarly, AI can be used to generate fake information, such as fake images or fake dialogues, through chatbots. These comments, generated using AI tools, could influence public opinion and thus their belief. Moreover, the fake dialogues generated by Chatbots can give a false impression of engagement, leading to less personal engagement and potential misunderstanding if the AI does not accurately interpret the users’ posts. In that sense, people with a less sophisticated understanding of AI could use these tools to generate content to boost particular political agenda. This could lead to an overwhelming amount of low-quality AI-generated content being produced. All these could result in a higher risk of misinformation, which could distort voters’ perception of certain candidates. Dr Fahmi, therefore, believed that there is a need to have a global and national regulatory body to regulate the use of AI in all aspects, including its usage during electoral campaigning.
The webinar was attended by 95 participants from the region and beyond. During the Q&A session, speakers discussed key topics such as the possibility of foreign state operators using AI to influence the upcoming 2024 elections, the lack of a policy outline for social media usage in the forthcoming elections, key challenges in regulating AI usage, Indonesia current level of awareness towards disinformation and the effects of civil society organizations in mitigating such effects. Lastly, a question was raised on the potentiality of Indonesia’s presidential candidates using TikTok as a form of outreach to the voters, given that this mode of communication was the main driving factor in attracting young Malaysian voters during their recent General elections. Dr Fahmi believed that there is a high chance that the presidential candidates in Indonesia would use this platform to drive engagement with the society. That being said, he also cautioned the need to regulate its usage, as well as making sure that the information posted online is credible to prevent misinformation.