Recently, several videos generated by artificial intelligence (AI) have been circulating on social media, depicting confrontations between police officers in New York City uniforms and federal immigration enforcement officers, with scenarios even showing the “arrest of ICE officers,” leading some netizens to mistakenly believe that the scenes are real.
These videos show instances of New York City police officers scolding, arresting, or directly confronting United States Immigration and Customs Enforcement (ICE) personnel, and the viewership on various platforms has been steadily increasing in recent times. Due to the sensitivity of the issues involved, there is considerable online attention and interest in these videos.
Experts point out that these videos are not labeled as AI-generated, but their authenticity can be discerned from details such as blurry subway signage text that is indecipherable or obvious inconsistencies in the actions of individuals and the logic of the scenes.
Both the New York City Police Department and the Department of Homeland Security have not responded to related inquiries, and there are no credible reports indicating that the New York City police have ever arrested federal immigration enforcement officers. In fact, the New York City police do not participate in civil immigration enforcement actions.
One video identified by experts as AI-manipulated shows multiple police officers in “NYPD” uniforms arresting a man in an “ICE” uniform on a busy subway platform, with surrounding passengers cheering and applauding. Another video depicts what appears to be an arrest taking place at Times Square, where police detain a man wearing military-style clothing with “ICE” written on it.
Lindsay Gorman, a former senior technology advisor in the Biden administration, pointed out that some of these videos have actually been circulating online for several months, but with the immigration enforcement issue heating up, they have been widely shared again.
Scott Brennen, director of the Technology Policy Center at New York University, expressed that while many may be able to recognize these videos as AI-generated, what is more concerning is the potential “secondary effects” that could undermine the public’s trust in all visual evidence.
“When you are in an environment where a large amount of content is generated by AI, people are more likely to question, or even deny, the reality of images, including videos that truly expose violence or major events,” Brennen said.
There is a high level of public interest in information concerning interactions between local and federal law enforcement agencies in the United States, but reliable information is relatively scarce, providing fertile ground for propaganda.
Currently, the identities of the creators of these videos are unclear. Gorman suggested that their motives could be diverse, ranging from simple artistic or political expression, symbolic creations opposing ICE, to deliberate manipulation of public opinion and exacerbation of political divisions.