Indonesians are deeply aware of the power of AI-generated content to manipulate public opinion – but those who think they’re safe may be the most vulnerable, according to new research from Luminate conducted by Ipsos.
Three in four (75%) believe AI–generated content has the potential to impact the general public’s political views, while 72% say it could affect their close friends and family. Even at a personal level, 63% acknowledge that AI-generated content could shape their own political views.
This suggests a broad awareness that AI-driven narratives have real-world consequences, even as individuals may underestimate their own susceptibility compared to those around them.
Those who think AI won’t affect them may struggle to detect it
Emergent research shows that AI-generated propaganda can be more effective than content produced by people. What’s less clear is whether it still works if people know it was created by AI.
Regardless, our survey found that significant numbers of Indonesians aren’t confident in their ability to identify whether social media content is AI-generated. A quarter (26%) admit they are not very confident, or not confident at all; 70% claim to be at least fairly confident.
Worryingly, among the 33% of Indonesians who believe AI-generated content has little to no impact on their political views, 42% admit they are not very/at all confident in identifying AI-generated content.
In a country of 200 million voters, this suggests a significant blind spot in Indonesia’s digital landscape: many people who assume they are immune to AI-driven narratives may, in fact, be the easiest to manipulate.
And while men and women in Indonesia report similar overall confidence in identifying AI-generated content (70% and 71%, respectively), a stark difference emerges in the level of confidence they express. Only 17% of women say they are “very confident” in identifying AI-generated content, compared to 30% of men. This suggests that women may be underestimating their own abilities or that men may be overestimating theirs.
Very high levels of social media consumption
Indonesia’s very high reliance on social media highlights the scale of these risks. Messaging apps and visually driven social media platforms capture a large share of people’s attention in Indonesia. A majority of people surveyed of all ages (21 to 65 years old) said they use Facebook, Instagram, and WhatsApp at least once per day.
WhatsApp dominates daily usage, with 94% of adults surveyed using it at least once a day, and nearly three-quarters checking it more than three times daily. Facebook and Instagram also see heavy traffic, with 70% of users logging in daily, while TikTok isn’t far behind at 64%.
These findings underscore how critical digital platforms may be in shaping political discourse -- for better or worse. With such high exposure to digital content, AI-generated narratives can spread quickly and take hold before users even realise they are engaging with them.
Indonesia as a case study in AI’s impact on democracy
At Luminate, we’ve been mapping attitudes toward AI, democracy, and social justice over the past 18 months around the world, including key countries in Latin America and Europe.
As the world’s third-largest democracy with exceptionally high digital penetration, Indonesia offers important insights into the ways tech platforms influence information ecosystems, particularly following a presidential election in which the winner used generative AI to sway votes – efforts that had wide reach and were likely effective at framing political discourse.
From algorithms to synthetic content, the true nature and scale of AI’s impact on democracy continues to be explored. In this setting, our findings reveal the extent to which even advanced democratic systems may not be ready for the AI wave.
The AI era demands collective intelligence
These findings reveal uncertainty among Indonesians about their ability to separate fact from fiction in an AI-driven information landscape and paints a picture of citizens in a large democratic nation who feel they are losing control over their shared reality.
This means investing in community-led digital resilience and awareness programmes, demanding transparent AI governance and safety audits, and ensuring that those on the frontlines of digital harm have the power to shape the policies and technologies that govern their lives.
The solution to AI-driven disinformation cannot come from tech companies alone. It must come from collective intelligence, drawing on the insights of civil society, frontline groups, marginalised communities, and experts who understand the lived realities of disinformation’s impact. If we are to build a more resilient digital space, we must source and scale the best ideas from those most affected.
Methodology
Ipsos interviewed a representative sample of 1000 adults aged 21-65 in Indonesia using an online methodology between 28th November and 6th December 2024.
Broad quota controls were set on age, gender, region and working status. The data has been weighted to the known population proportions for gender by age, region, and working status.
All design and reporting was conducted by Luminate with Ipsos carrying out the data collection. The survey was conducted in Indonesian (Bahasa).