Experts Warn of Misinformation Risks Posed by Google's Veo 3 in Sweden

In July 2025, Google unveiled its AI video tool Veo 3 in Sweden, which enables users to generate realistic news videos quickly. However, experts have raised alarms about its potential to spread misinformation and incite social unrest. According to Carl Heath, a researcher in digital resilience, the technology greatly simplifies video production, allowing users to create an eight-second video with lip-synced dialogue in mere seconds by providing a text prompt.

In tests conducted by SVT, while many videos were somewhat unrealistic, one clip falsely reported an Ebola outbreak in Sweden, demonstrating the tool's potential for harmful misinformation. Andreas Jahrehorn Önnerfors, a professor of intellectual history, warned that AI-generated content could easily become a trigger for violence, stating, "A video like this could ignite a powder keg." This sentiment is echoed by Heath, who noted that malicious entities could exploit Veo 3 to manipulate public sentiment or influence political scenarios.

To mitigate misuse, Google has implemented watermarks on AI-generated videos aimed at making them identifiable. Nonetheless, the detection technology, SynthID, is not yet widely available, limiting its effectiveness. During SVT's testing, while some violent content was successfully prevented from being generated, the system could still be bypassed, raising further concerns about the robustness of safeguards.

Despite these risks, Önnerfors suggested potential benefits in using AI-generated videos to provide news coverage in areas lacking journalistic resources, indicating a dual-edged nature of such technology. While Veo 3 could enhance news reporting capabilities, it equally poses significant risks regarding the spread of disinformation and the shaping of public perceptions.