The use of AI in political communication affects trust even if it is not detected

The use of artificial intelligence in political communication has been shown to affect public trust even when the content is not recognized as synthetic, according to a recent study conducted by Civic Compass. The research analyzed how audiences react to politically generated content, both original and AI-generated, and found that while the origin of the content significantly influences perceptions of credibility, truthfulness, emotional connection, and trust, many participants failed to consistently identify the AI-generated material. In a rapidly evolving media landscape, where social networks and video platforms have surpassed traditional news websites and apps as primary sources of information, the role of AI in shaping public discourse has become increasingly relevant. According to the Digital News Report 2026 from the Reuters Institute for the Study of Journalism, this shift underscores a growing challenge for democratic societies, particularly with the rise of AI-driven technologies capable of generating synthetic content such as images, videos, and audio. These tools not only question the authenticity of information but also raise concerns about how citizens evaluate the reliability of the messages they receive. Katie Harbath’s recent study highlights the widespread adoption of AI in political campaigns. It reveals that 87% of campaign professionals use AI tools daily or several times a week, and 74% believe these technologies will be essential for maintaining competitiveness in future elections. This trend suggests that AI is becoming an integral part of modern political strategy, further complicating the relationship between information dissemination and public perception. To explore these dynamics, Civic Compass designed an experiment examining how the source of political content affects audience responses. Participants were exposed to both original and AI-generated versions of a speech by Argentine politician Javier Milei, delivered outside the 2025 election calendar. The study aimed to understand how the origin of the content, whether human-generated or AI-produced, influences perceptions of credibility, trust, and emotional engagement. Researchers also sought to determine the criteria individuals use to assess the authenticity and truthfulness of political messages. The findings revealed that AI-generated content systematically undermines key aspects of public trust. On average, AI-generated content reduced perceived credibility by 9.2 percentage points, truthfulness by 5.2 percentage points, emotional connection by 3.9 percentage points, and trust in the person or organization sharing the message by 7 percentage points. Despite these measurable differences, participants often did not consistently recognize the AI-generated nature of the content. This suggests that the influence of AI extends beyond just altering the content itself, it may also shape how people interpret and evaluate the information they consume. Furthermore, the study identified variations in how different demographic groups responded to AI-generated content. While some segments of the population were more likely to question the authenticity of AI-generated messages, others showed less awareness of the technology's role in content creation. These disparities highlight the need for greater public understanding of AI's capabilities and limitations in the context of political communication. As AI continues to play a larger role in political messaging, its impact on public trust and civic engagement remains a critical area of concern. The implications of these findings suggest that the increasing reliance on AI in political strategies could lead to broader challenges in maintaining informed citizenry and fostering trust in democratic institutions. Future research will need to address how these effects evolve over time and whether new measures can help mitigate the risks associated with AI-generated political content.