AI-generated violent extremist content
Use of generative artificial intelligence to disseminate extremist content is a field of research that examines the application of generative AI technologies (such as Viggle AI, HeyGen, and Reface) to create and spread propaganda that glorifies violence, terrorist attacks, and mass killings. In 2025–2026, experts observed a trend toward using readily available AI tools to produce video content that normalizes violence and targets young audiences on platforms such as TikTok, Instagram, and other social media sites.[1] [2]
Forms of extremist AI-generated content
[edit ]Viggle AI and the "Rampage dance" trend
[edit ]Viggle AI is Chinese artificial intelligence-based image-to-video platform. It allows users to place people or fictional characters from images into reference videos and transform static images into animated videos, including an AI Dance Generator.[3] Viggle AI was launched in March 2024 by founder and CEO Chu Hang, a former employee of Google, Nvidia, Facebook, and Autodesk. By February 2026, the platform had surpassed 40 million users.[4] [5]
Global Network on Extremism and Technology (GNET) researchers have identified widespread use of Viggle AI to create videos of the "rampage dance" trend, in which well-known mass murderers are depicted dancing against the backdrop of the sites of their attacks, with a reference to their victims—usually the number of people killed—along with an emoji indicating their religious affiliation. In Tarrant’s case, "-51" (the number of people killed in Christchurch) also typically appears on the screen, followed by the "☪️" emoji, which is used to glorify both the number of victims of the attack and their religion in an Islamophobic manner.[1]
Members of this True Crime Community (TCC), who are obsessed with serial killers and mass murderers, actively use Viggle AI to create videos glorifying criminals. Among the most frequently depicted figures are:
According to researchers, such videos do not so much directly incite violence as they contribute to the normalization of violence, transforming it into an aestheticized act and a form of entertainment.[1] Most of the videos found depicted gunmen dancing and were used on TikTok as part of the so-called "Rampage dance" trend.[1]
Salafi-jihadist communities
[edit ]In Salafi-jihadist communities, Viggle AI is used to glorify well-known terrorist attacks. Among the figures depicted are "Jihadi John", participants in the September 11 attacks, the 2019 suicide bombings in Sri Lanka, the attack on Crocus City Hall, the attacks on Bondi Beach, and the Pulse nightclub shooting. The videos contain the number of victims in each attack, and the figures themselves are depicted striking aggressive poses while performing a viral dance trend, similar to the aforementioned "Rampage dance" trend.[1]
AI-generated fruit and vegetable gore content
[edit ]Researchers have identified a distinct genre—AI slop content—in which anthropomorphic fruits and vegetables are subjected to violence, torture, and hyper-realistic killings. This genre, referred to as "AI fruit/vegetable gore," uses aesthetics appealing to children (bright colors, Pixar-style animation) to mask its violent content. The videos contain references to various extremist groups and ideologies, such as videos depicting the violent methods of cartels, reenactments of scenes from ISIS propaganda videos and the use of hashtags typical of ISIS supporters, references to the neo-Nazi organization Atomwaffen Division and supremacist slogans, and references to the "No Lives Matter" subculture. In some videos, the character "Carrot Caliph" declares that "the Carrot State will rise again," mimicking the slogan "the Islamic State will rise again".[6]
Post-organizational extremism
[edit ]The dissemination of extremist propaganda using AI has become a central element of a new reality that experts refer to as "post-organizational extremism." The key idea is that extremist ideology can now exist and spread independently of any formal organizational structure. AI acts as an amplifying factor, enabling individuals ("lone wolves") inspired by a particular worldview to create and disseminate propaganda without direct ties to any group. This makes traditional counterterrorism approaches, which focus on combating specific organizations, less effective.[1] Scholars are now talking about an era of "ideological fluidity," in which old distinctions between far-right, jihadist, and nihilist factions are breaking down.[7]
Generative AI technologies make it possible to create and disseminate propaganda in easily shareable formats on mainstream platforms, while preserving the anonymity of both content creators and consumers.[1] As of March 20, 2026, the HeyGen platform had, according to its own claims, generated 113,457,133 deepfake videos.[1]
2026 Tumbler Ridge shooting
[edit ]The Tumbler Ridge attack that took place on February 10, 2026, became an example of how AI can serve as a medium for radicalization.[8] [9] Eighteen-year-old Jesse Van Rootselaar shot and killed her 39-year-old mother and 11-year-old stepbrother at home, then went to a local high school, where she killed five students aged 12–13 and one teacher, before taking her own life. Twenty-five people were injured. It became one of the deadliest school shootings in Canadian history. Van Rootselaar, who was 17 at the time, had spent several days describing various scenarios involving gun violence to ChatGPT. OpenAI’s automated abuse detection system logged these conversations.
About 12 members of OpenAI’s security team manually reviewed the correspondence and identified it as indicating an "imminent risk of serious harm to others." They recommended notifying Canadian law enforcement.[10] OpenAI’s management rejected this recommendation. The company suspended Van Rootselaar’s account for violating its terms of service but did not contact law enforcement because, in management’s assessment, the conversations did not meet the threshold of a "credible and imminent" threat.[11]
After being blocked, Van Rootselaar created a second account and continued her planning via ChatGPT; this account was not detected by OpenAI until after the attack took place.[10] [11]
See also
[edit ]References
[edit ]- 1 2 3 4 5 6 7 8 Lucini, Silvano Rizieri (2026年02月26日). "An 'Ode to Violence': Extremist Exploitation of Viggle AI". GNET. Retrieved 2026年08月25日.
- ↑ "AI and Extremist Propaganda: An Assessment".
- ↑ Carry, Owen (2026年03月11日). "Teens Are Using AI-Fueled 'Slander Pages' to Mock Their Teachers". Wired. ISSN 1059-1028 . Retrieved 2026年08月24日.
- ↑ "What Is Viggle AI And How To Use It? - Dataconomy". 2024年04月18日. Retrieved 2026年09月18日.
- ↑ Lucini, Silvano Rizieri (2026年02月26日). "An 'Ode to Violence': Extremist Exploitation of Viggle AI". GNET. Retrieved 2026年08月24日.
- ↑ Pugnana, Alessandra (2026年08月10日). "Fruit, Vegetable, and Brain Rot: The Use of Generative AI to Produce Violence-Based Viral Trends". GNET. Retrieved 2026年08月25日.
- ↑ "Gen-Z and Reconfiguration of the Violent Extremist Landscape". MP-IDSA. Retrieved 2026年08月25日.
- ↑ Herberhold, Allizandra (2026年05月01日). "The 'True Crime Community Stoplight Model': Understanding Digital Pathways and Tumbler Ridge". GNET. Retrieved 2026年08月25日.
- ↑ Varga, Michael (2026年07月09日). "The Counterterrorism Blind Spot: AI as a Radicalisation Environment". GNET. Retrieved 2026年08月25日.
- 1 2 Duffy, Clare (2026年04月29日). "Families of Tumbler Ridge shooting victims sue OpenAI and CEO Sam Altman | CNN Business". CNN. Retrieved 2026年08月25日.
- 1 2 "OpenAI vows safety policy changes after Tumbler Ridge shooting". www.bbc.com. 2026年02月27日. Retrieved 2026年08月25日.
Further reading
[edit ]- "An Ode to Violence: Extremist Exploitation of Viggle AI". Global Network on Extremism and Technology (GNET). February 26, 2026.
- "Fruit, Vegetable and Brain Rot: The Use of Generative AI to Produce Violence-Based Viral Trends". Global Network on Extremism and Technology (GNET). August 10, 2026.
- "The True Crime Community Stoplight Model: Understanding Digital Pathways and Tumbler Ridge". Global Network on Extremism and Technology (GNET). May 1, 2026.
- Kumar, A. (August 25, 2026). "AI and Extremist Propaganda: An Assessment". Institute for Defence Studies and Analyses (IDSA).