AI-Driven Propaganda Raises Concern as Synthetic Media Detection Becomes More Challenging

As artificial intelligence technology advances, experts are raising alarms about a fundamental shift in propaganda methods from broad messaging to highly personalized psychological manipulation targeting individual vulnerabilities.

In response to growing concerns, X (formerly Twitter) has updated its creator policy specifically addressing AI-generated war videos. Under the new guidelines, creators who fail to clearly label synthetic war content as artificial face serious consequences, including loss of monetization privileges and potential permanent expulsion from the platform’s revenue sharing program.

“Historically, propaganda was amplified through television debates, newspapers, or mass forwards on messaging platforms. AI has transformed propaganda from a loud broadcast into a personalised whisper,” said Kartik Gupta, Instructor of AI and Machine Learning at Newton School of Technology.

The evolution of AI systems has enabled unprecedented capabilities for analyzing personal data. Modern algorithms can process behavioral patterns, linguistic preferences, browsing history and social engagement metrics to identify specific vulnerabilities in individuals. This allows for the creation of tailored narratives precisely calibrated to resonate with a person’s cultural background, religious beliefs, or political preferences.

Making matters worse, the quality of synthetic media has improved dramatically. A biometric study conducted in late 2025 revealed that only a small percentage of participants could accurately distinguish between authentic content and AI-generated material, highlighting how AI-driven propaganda can erode trust in information before manipulation is even detected.

Social media platforms have become the primary vectors for distributing AI-generated misinformation, according to Atul Rai, Co-founder and CEO of Staqu Technologies. “Deepfakes and synthetic media gain traction because platform algorithms prioritize engagement, allowing manipulated content to reach large audiences,” Rai explained.

He emphasized that platforms must leverage their technological capabilities to implement advanced AI systems capable of identifying manipulated audio, synthetic video, and coordinated bot networks designed to amplify propaganda. Additionally, stronger governance frameworks are needed, including rapid response protocols during geopolitical crises, transparent labeling of AI-generated content, and partnerships with fact-checking organizations.

Responsibility extends beyond platforms alone, noted Kaushal Bheda, Director at Pelorus Technology. Content creators who deliberately generate disinformation or propaganda bear direct accountability for resulting harm. Meanwhile, platform developers and operators must implement effective preventative measures and respond quickly to law enforcement requests.

When authorities identify active harm campaigns, delays in providing data or suspending accounts allow damage to compound. “Immediate cooperation with investigations, expedited legal compliance, and proactive intelligence sharing are non-negotiable responsibilities for platforms operating at global scale,” Bheda said.

The verification of digital content is becoming increasingly critical as society enters an era where images, voices, and videos can no longer be presumed authentic by default. Gupta noted that verification must transition from an individual responsibility to a systemic approach. This requires governments, platforms, and educational institutions to develop stronger early-warning systems, authentication protocols, and rapid response frameworks, particularly during high-risk periods like elections or natural disasters.

“There may be difficult debates ahead around temporary amplification controls during national emergencies. While controversial, such measures reflect a broader tension between open digital ecosystems and public safety,” Gupta added.

Garry Singh, President of IIRIS, highlighted that while major platforms have implemented mechanisms to identify AI-generated propaganda, a significant challenge remains in the speed of content moderation. “The mechanism to address complaints and remove bad content is slow, causing concerns of spreading before the content can be taken down,” Singh explained.

The potential risks extend far beyond misinformation, encompassing threats to life safety, financial security, resource allocation, and public opinion formation. As AI-generated content becomes increasingly sophisticated, the line between reality and manipulation continues to blur, creating unprecedented challenges for social platforms, regulators, and society at large.