How is public perception of social media shifting globally?
Public concern regarding the impact of social media on democratic stability is rising significantly in many developed nations. A recent Pew Research Center study covering 37 countries indicates that residents of high-income nations—including the US, Australia, Canada, France, Germany, the UK, and the Netherlands—are increasingly likely to believe that social platforms are detrimental to democratic life. In at least half of these seven nations, more than 50% of citizens hold this negative view.
The trend toward skepticism is particularly pronounced in certain European countries. For example, in Germany, the percentage of people viewing social media as harmful to democracy has jumped from 41% to 61%. Similarly, in the Netherlands, the figure rose from 54% to 70%, while in France, it climbed from 51% to 65%. Japan has also seen a notable increase, moving from 23% to 43%.
However, the global landscape is not uniform. In countries such as Ghana, Kenya, Nigeria, the Philippines, and Thailand, approximately three out of four people view social media positively. Furthermore, in high-income nations like Israel and Singapore, only 22% of citizens believe social media harms democracy. This divergence suggests that the perceived threat is closely tied to regional political contexts and digital maturity.
Demographic divides in digital sentiment
Age plays a critical role in how individuals perceive the democratic value of social platforms. In the United Kingdom, 60% of those aged 18-34 view social media as beneficial to democracy, compared to just 31% of the population aged 50 and over. A similar pattern is observed in the United States, where 44% of younger users hold a positive view, compared to 28% of older demographics.
What is the role of AI in creating 'troll factories'?
Artificial intelligence is dramatically lowering the cost and increasing the speed of producing deceptive political narratives. Experts suggest that AI does not necessarily create new types of manipulation, but rather automates and accelerates the existing 'troll factory' model. Instead of requiring a large team of human actors to craft and spread a story, a single user can now provide a prompt and distribute content almost instantaneously.
Katie Harbath, a former Meta public policy expert, explains that AI introduces a new level of complexity to the misinformation landscape. She notes that while AI allows for the creation of much larger volumes of content that is more personalized, faster, and cheaper to produce, the primary distribution mechanism remains traditional social media platforms.
This efficiency enables the mass production of highly targeted content. As Rossine Fallorina from a Philippine AI think tank points out, the technology essentially mimics human-led disinformation campaigns but at a scale previously unimaginable. This capability allows bad actors to flood the digital space with manufactured narratives designed to exploit existing societal fractures.
How is AI being used for election manipulation?
Documented instances of AI-driven political interference demonstrate the tangible risks these technologies pose to electoral integrity. In Malaysia, an attempt to manipulate the electorate was uncovered involving the Claude AI model. This operation utilized a commercial platform to target voters across 222 electoral districts through a network of approximately 1,000 accounts on X, a fake news website, and a series of forged documents.
Anthropic, the developer of Claude, recently revealed that its AI was being used in this attempt to influence political outcomes. While the experiment ultimately failed due to intervention by Anthropic, the company's September report detailed how various actors—including governments, pro-government foundations, state media, and lobbyists—have used the tool to create fake social media profiles and news sites across six continents.
The threat is not limited to Southeast Asia. In Israel, ahead of the October 27 elections, reports indicate that even politicians and ministers are utilizing AI-generated videos to attack opponents. According to the Ze Be AI initiative from the Open University of Israel, approximately one in eight election-related posts has been created or modified using AI. Despite new laws requiring the labeling of AI content, regulators struggle to keep pace with the speed of digital dissemination.
Why does AI-generated content drive social polarization?
AI-generated content is specifically designed to exploit pre-existing political anxieties and cultural grievances to fuel social division. On platforms like TikTok, experts have identified a surge in AI-generated videos depicting fake protests across Europe. These videos, some of which have garnered over 300,000 views, often depict large-scale demonstrations against migration or current governments in countries like Portugal, France, Germany, Ireland, Italy, and the UK.
In many cases, these protests never actually occurred. One specific example involved a video showing a journalist holding a microphone with an RTP logo; however, the Portuguese state broadcaster confirmed the footage was entirely synthetic and the individual pictured was not an employee. Such content is engineered to trigger emotional responses and reinforce existing biases.
Dr. Ben Loh, a media lecturer at Monash University in Malaysia, observes that people with deeply rooted prejudices often use AI content to validate what they already believe to be true. Bruna Martins dos Santos from the international organization WITNESS emphasizes that these synthetic images are particularly effective because they directly tap into existing societal tensions, making them highly potent tools for those seeking to undermine democratic dialogue.
Frequently asked questions
Does AI make misinformation more dangerous?
Yes, AI makes misinformation significantly more dangerous by increasing the volume, speed, and personalization of deceptive content. It allows a single actor to operate at the scale of a large organization, creating 'troll factories' that can flood social media with tailored narratives designed to manipulate specific voter groups.
Can social media be both good and bad for democracy?
Most people believe social media has a dual nature. While 79% of surveyed individuals believe the internet facilitates manipulation via false information and 68% see it as a cause of political division, many also acknowledge its benefits. For instance, 77% believe it makes people better informed, and 60% feel it provides a more meaningful political voice.
How are governments responding to AI in elections?
Governments are attempting to implement regulations, such as Israel's July law requiring the labeling of AI-generated election materials. However, researchers warn that enforcement is often too slow. By the time a deceptive video is identified and removed, the psychological and political damage to the electorate has often already been done.
Is AI-generated content always fake?
Not necessarily, but it is increasingly used to create 'deepfakes' or synthetic media that look real. Even when politicians use AI with proper disclosures, the technology is frequently used by anonymous actors to create entirely fabricated events, such as the fake European protests seen on TikTok, to stir social unrest.
What is the main goal of AI-driven polarization?
The primary goal is to manipulate public opinion and deepen social and political divisions. By using AI to target specific cultural grievances and existing biases, bad actors can create a sense of crisis or widespread discontent, which undermines the stability of democratic discourse and trust in institutions.
Key takeaways
- AI enables the creation of cheap, high-volume 'troll factories' that automate political disinformation.
- Public skepticism toward social media is rising in wealthy nations like Germany, France, and the Netherlands.
- The Claude AI model was recently identified in an attempt to manipulate Malaysian elections.
- AI-generated fake protests on TikTok are being used to fuel social polarization across Europe.
- Regulators struggle to keep pace with AI, as the damage from misinformation often occurs before removal.
Conclusion
The intersection of artificial intelligence and social media represents a profound challenge to modern democratic stability. As AI lowers the barriers to large-scale manipulation, the ability of bad actors to manufacture consensus or incite division through synthetic content grows exponentially. While digital platforms continue to offer tools for political engagement and information sharing, the rise of automated 'troll factories' threatens to drown out authentic discourse. Addressing this threat will require not only faster regulatory responses but also a global effort to enhance digital literacy and protect the integrity of the information ecosystem.
