Last week, the release of OpenAI’s Sora 2, an advanced AI video generation tool, triggered a new wave of hyper-realistic synthetic videos across social media platforms. While the technology has sparked fresh debates about intellectual property and digital authenticity, it’s also fueled a more disturbing trend: the proliferation of AI-generated protest videos depicting violence that never happened. These clips, often indistinguishable at first glance from real footage, are now shaping political narratives and sparking confusion—and outrage—both online and off.
On October 7, 2025, one such AI-generated video went viral on Instagram, racking up over 40 million views. The video features a protester, rendered by Sora 2, shouting "What’s your name, soldier?" at a camouflaged figure. The tension builds until the AI soldier responds with a blast of orange pepper spray, retorting, “Sergeant Pepper.” The video’s overlay reads, “wait for it,” inviting viewers to witness the manufactured confrontation. According to reporting from Gizmodo, the video quickly spread across TikTok and X (formerly Twitter), where many users failed to recognize it as artificial. Actor James Woods, a well-known supporter of President Donald Trump, praised the video on X, calling it “Haiku-level brilliant.”
Another AI-generated video, this time featuring protesters chanting a racist riff on the familiar slogan “no justice, no peace,” has also gained traction, amassing over 1.5 million views on X. In this clip, the AI protesters are again sprayed with a chemical agent. The video’s caption, “Lmfao, that was beautiful and before you ask I voted 100 for this!!” echoes a phrase that has become popular among Trump’s far-right supporters, often used to celebrate perceived victories over political adversaries.
The Instagram version of the video goes further, labeling the protesters as “clowns” and using the acronym “FAFO”—short for “fuck around and find out.” Some social media users have tried to point out the AI origins, but the accounts sharing the clips often double down, insisting the events are real. Despite the prominent Sora watermark, many viewers remain convinced, highlighting just how blurry the line between reality and fabrication has become in the age of generative AI.
As BBC Verify Live reported on October 8, 2025, their team has been actively working to debunk and contextualize these viral protest videos. The BBC used a likely satirical AI-generated protest clip to demonstrate techniques for spotting fakes, reminding viewers to remain skeptical of unverified content circulating online. Their analysis extended to claims about Portland police confronting immigration agents—claims that proved to be misleading or outright false after close scrutiny.
But while AI-generated videos dominate the conversation, there are, tragically, plenty of real incidents of violence against protesters to document. In Chicago last month, Rev. David Black, a Presbyterian pastor, was shot in the head with pepper balls by ICE agents while peacefully protesting outside a detention facility. According to Mediaite, Black is now a plaintiff in a lawsuit filed by the American Civil Liberties Union (ACLU) of Illinois, which accuses ICE of using “violent force” on peaceful demonstrators. Black himself posted videos of the incident, stating, “This video shows ICE officers shoving me and sprayed me in the face with a large quantity of chemical weapons. I was standing there incapacitated and vulnerable from a barrage of chemical pellets they had shot me with moments earlier. ICE gave no orders or warnings. All around me, they were slamming peaceful protesters into fences and into the ground, and moving in formation to isolate and assault anyone who was vulnerable.”
Black claimed he was shot “seven times,” including in the head, with “chemical pellets.” He also wrote, “This is how the Trump administration treats the public witness of Christian faith. This and worse has been happening every day in Broadview.” The ACLU’s lawsuit, filed on October 8, 2025, names President Donald Trump, Acting ICE Director Todd Lyons, Attorney General Pam Bondi, Commissioner Gregory Bovino, and Kristi Noem as defendants. The suit alleges that, in recent weeks, the Trump Administration “has sent federal officials and agents to brutally suppress free speech at the site through intentional and escalating violence, including the dangerous and indiscriminate use of near-lethal weapons such as tear gas, rubber bullets, pepper-balls, flash grenades, and other unwarranted and disproportionate tactics.”
Other real-world incidents have been caught on camera as well. On October 3, 2025, a video surfaced from Portland, Oregon, showing a woman speaking calmly with police officers before one suddenly pepper-sprays her in the face. According to Gizmodo, the woman was left in visible pain and required assistance. These genuine acts of violence stand in stark contrast to the AI-generated clips, which often serve to reinforce political talking points rather than document reality.
The context for these protest videos—both real and synthetic—is President Trump’s recent attempts to deploy National Guard troops to cities like Portland and Chicago. Although court orders have delayed these deployments, the presence of federal agents, particularly from ICE, remains a source of fear and tension in many communities. Trump has repeatedly claimed that these cities are “overrun with crime” and that his actions are necessary to restore law and order. However, as Gizmodo points out, violent crime is near a 50-year low, contradicting the president’s narrative and raising questions about the motivations behind the escalated federal response.
The economic impact of these crackdowns is also being felt. Restaurant owners in Chicago have compared the recent downturn in business to the early days of the COVID-19 pandemic, attributing the loss in part to the heavy federal presence and the unrest it brings. This, in turn, has created fertile ground for the spread of fake protest videos, which some see as a way to justify or rationalize the ongoing crackdowns.
Even President Trump himself has been caught up in the confusion between real and artificial. He recently shared an AI-generated video on his Truth Social account promoting a so-called “med bed” that claims to cure all diseases. The video was later deleted, but the incident left many observers wondering whether Trump recognized the clip as a fabrication. Back in April, he insisted a photoshopped image of “MS-13” tattooed on a protester’s hand was authentic, further blurring the lines between reality and fiction for his followers.
The consequences of this new media landscape are profound. As BBC Verify Live’s coverage demonstrates, efforts to fact-check and debunk misleading content are more important than ever, but the sheer volume and sophistication of AI-generated media make the task daunting. Meanwhile, real acts of violence against peaceful protesters—like those experienced by Rev. David Black and others—can be overshadowed or distorted by the flood of synthetic imagery, complicating public understanding of what’s actually happening on America’s streets.
As AI tools like Sora 2 become more accessible and powerful, the challenge for journalists, activists, and everyday citizens will be to distinguish between the virtual and the real. The stakes, as recent events have shown, could not be higher.
