AI, Deepfakes, and Indonesia's 2024 Election
Welcome to my newsletter on data and AI in Asia! Here I talk about things happening in the AI and data space in the region. I mostly touch upon things in China and Indonesia but will dive into developments in other countries as well.
In this newsletter, I take a break from exploring the regional AI industry to look at how AI is impacting the democratic process in one of the countries that go to the ballot boxes this year, Indonesia.
?
A few takeaways:
?
?
?2024 is setting a record for the greatest number of people in countries with nationwide elections. More than four billion are eligible to vote this year. That’s half of the world’s population.
In Asia, people in the two most populous countries India and Indonesia are going to the polls. Indonesia is a relatively young democracy. It is the largest economy and most populous country in Southeast Asia. Indonesia also has the largest Muslim population in the world. On February 14, three candidates square off in the country’s presidential election.
The country’s population is young, the country’s digital economy is the largest in Southeast Asia, and social media plays an important role in everyday life. Most Indonesians take in news on one or several social media platforms.
This combined with technological advancements has resulted in increasing worries about the potential adverse effects that technologies such as AI may have on the democratic process in the country.
AI-generated deepfakes trying to influence voters have been reported and discussed widely online over the last year.
One example of this is a video of Indonesia’s President Jokowi that started circulating a few months ago – it showed him giving a speech in fluent Chinese. This suggested to viewers that his government is controlled or at least significantly influenced by China.
?
Other examples of deepfakes are TikTok videos of two of the Indonesian presidential candidates, Prabowo Subianto and Anies Baswedan, speaking in fluent Arabic.
Recently, a deepfake with the late Indonesian authoritarian leader Suharto urging voters to vote for the political party Golkar was posted online. A senior member of Golkar posted and pinned it on X.
?
?
Activists have called on the country’s General Elections Commission (KPU) to regulate the use of AI in campaigning. They worry that it may end up threatening the integrity and fairness of the democratic process.
AI has also been used for other things related to the election. For example, the team behind Prabowo Subianto has generated images of him and his running mate Gibran Rakabuming Raka as cute kids. They can be seen everywhere on billboards and TV ads.
领英推荐
For obvious reasons, AI-generated content such as the ones of Prabowo and Gibran as cute kids created and disseminated by their team can hardly be seen as the same or even similar to deepfakes. It is, however, an example of how AI-derived content is playing an increasingly important role in content creation and messaging.
Given the rapid development of AI, deepfakes and other forms of AI-generated misinformation can end up having a major influence on individuals voting behavior. The impact is likely higher if information and digital literacy levels are low.
In May 2023, Microsoft President Brad Smith gave a speech during which he stated that his biggest concern around AI was deep fakes, realistic looking but false content.
Similarly, the recently released World Economic Forum’s “Global Risks Report 2024” ranked AI-generated misinformation and disinformation as the biggest global risk in 2024, ahead of climate change, war, and economic weakness.?
Indonesia introduced its national strategy for AI back in 2020. However, it did not focus on information and online content. The General Elections Commission (KPU) has also said that it will not regulate AI campaigning during the 2024 election process.
As noted by scholar Nurrianti Jalli , there is a significant need for a policy that focuses on the information realm so that the positives of AI can be promoted while the potential negatives are curtailed.
AI-generated content is not going to go away. I believe we can expect an increasing share of all output online to be generated by AI in one form or another, regardless of whether it is text, image, or video.
A lot of AI-generated content can still be distinguished by viewers. It is, however, increasingly challenging for average citizens to differentiate between real and fake.
The challenge will be to regulate it constructively and efficiently without limiting people’s right to express themselves.
?
?
?
?
?
?
A few select readings if you’re interested in diving deeper into the 2024 elections in Indonesia and elsewhere as well as the use of AI in the electoral process:
?
?