Swamp Slop and Fake Moats: On the AI Mediation of Alligator Alcatraz
Diana Flores Ruiz / University of Washington

Fig 1. Author’s screenshot of an AI Image of alligators in ICE hats from Offical DHS X Account

Ten days after the concept of “Alligator Alcatraz” was put forward by Florida Attorney General James Uthmeier in June of 2025, the Department of Homeland Security (DHS) shared an AI-generated image of alligators wearing Immigration and Customs Enforcement (ICE) hats. The official post captions the gator guards in front of a prison fence and watch tower simply: “Coming soon!”. The pithy, comical tone of the post warranted follow-up news coverage affirming it as an official post as well as debates on its humorous tone. While the incredulity of such a flippant tone settled in as the latest indication of a meme-ified approach to normalizing a mass detention and deportation regime, the AI image literalized a fantasy about the natural capacities of alligators as predators and swamp ecologies as deterrents from incarcerated immigrants escaping the proposed facility. The cartoonish premise of the image catalyzed predictably polarizing discussions about the facility making the image’s mocking tone inseparable from its resulting celebration and outrage. But it was not the only popular AI-generated image depicting Alligator Alcatraz circulated that same day. 

Minutes after DHS’s post on the social media platform X, users shared an image of the facility surrounded by a moat, upon closer inspection, teeming with alligators. One such X user, a one-time Republican Congressional candidate for Nevada, includes in his caption: “WILD: Florida’s new detention site is now fully enclosed by gator-infested waters…No fences needed when nature handles security.” The click-bait appeal of his comments mixes proposed capacity details with blatant lies about the current state of the facility. He expresses himself through what I have described elsewhere as ecofascist aesthetics, or the idealized organization of nature according to authoritarian expelling of racialized others. Fact checkers found a post containing the moat image from the day prior to DHS’s announcement and showed how by the end of July, the image had been used in news articles off of social media.  

Fig. 2 Author’s screenshot of an X post from Tony Lane containing an AI image of the facility with an alligator-filled moat surrounding it

As the facility’s fate hangs in the balance of lawsuits stalled by the current government shutdown, I pause to consider how AI-generated images have shaped the mediation of arguments to close the facility as well as rallying cries to keep it open. I begin by sketching out two broad categories of AI images that have emerged in the months since the breaking news of its construction. I then examine some of the interplay between these categories, which allows me to show a more complex, effectively charged media landscape that exceeds simple truth claims about “fake” AI imagery.  

On the one hand, the image of alligators sporting ICE hats is emblematic of a larger trend mobilizing the breadth of emergent AI-generating platforms to literalize fantastical scenarios based on the image prompt of “Alligator Alcatraz.” I call this category swamp slop. Further examples include more propaganda posted on official White House social media, pro-prison memes in the formats of movie posters, and a slew of parodic video skits adapting the trope of reptilian weaponization into different sub-genres, such as Gator Gitmo: Miami ICE. Through self-aggrandizing mockery of detained people, swamp slop materializes the cruel undercurrents of the growing scale and punitive conditions of ICE detention and deportation. 

On the other hand, the less obviously AI-generated aerial view of the moat-surrounded facility represents media closer to what we understand as deepfakes in their realist aesthetics and ability to spread misinformation. The plausibility of scenarios depicted in these images yields a different valence of facticity and fuels their circulation through an expository, quasi-journalistic mode. This category includes a viral AI video that purports to show rising floodwaters flowing into caged blocks of bunk beds. This more subtle use of AI operates within the bounds of genuine risks and harms of the facility’s location, showing the consequences that environmental and anti-carceral activists have been warning about. As the Miami New Times reported, this video quickly gained nearly 10 million views and 10,000 re-posts, further extending its algorithmic reach onto other platforms.  

Fig 3. Author’s screenshot of an X post of an AI Video showing the flooding of the ICE facility

Despite their differences, these two categories of AI images circulate on the same platforms. Both categories emerge from the same charged political atmosphere and profit-driven media conditions, in which click-hungry users are trigger-happy in their generation of newly accessible tiers of AI imagery that perform well on social media. They are not mutually exclusive, often overlapping in animated videos attempting to realistically visualize the eco-carceral conditions of the facility. Given the historical lack of access to the inside of immigration detention centers, the ecofascist cultural fantasies that shape swamp slop create an outsized shadow over the visual iconography and extreme narratives around Alligator Alcatraz.  

To return to the fake flood video discussed above, its viral spread illuminated the dangers inherent in the facility’s remote wetlands location and powerfully illustrated the vulnerability of people incarcerated in it. In the process of fact-checking the video’s veracity, users consulted X’s embedded AI chatbot Grok, which incorrectly stated that the video was authentic and falsely attributed it to the reporter Jason Delgado. Even with the additional user traffic from Grok’s assessment, Delgado’s post featuring videos and photos of water seeping into the facility received, as of this writing, only three percent of the number of views as the AI flood video.

Fig.4 Author’s screenshot of X post from Jason Delgado about the actual event of water seeping into the facility which chat bots have been confusing with the AI flood video

I quantify this disparity to contextualize the stark difference between factual and AI-generated footage of the same phenomenon at Alligator Alcatraz. While arguments about the attention economy are relevant, I am more interested in the aesthetic and receptive effects of such an AI-saturated media landscape. What comes next? 

The effects of swamp slop extend beyond AI generated images, as in the case of the viral video that repurposed old news footage of an alligator removal to illustrate the claim that the largest alligators and most dangerous snakes were being selected for Alligator Alcatraz. Spectacles of misinformation continue to serve opposing perspectives on the facility in different capacities, such as allegory or satire, but the overwhelming majority of high-impression media depicting Alligator Alcatraz primarily concerns questions of veracity. 

The persistent pursuit of affirming or denying the veracity of potentially AI-generated images redirects multifaceted arguments about the facility, such as the interconnections between climate justice and migrant justice, into a true/false binary of corroborating or debunking specific hoaxes. Indeed, a DHS press release from August 2025 responds to multiple “Alligator Alcatraz hoaxes,” shutting down claims wholesale, rather than engaging in the degrees of harms reported from inside or predicted by experts. What is at stake in the AI mediation of the facility is a sustained distraction from the dangerous environmental and carceral conditions in favor of culturally reifying a new state formation of the federal detention and deportation regime. 

When deepfakes and swamp slop are so thoroughly embedded in news and popular discourse about Alligator Alcatraz, they normalize the underlying premise of exaggerated and far-fetched conditions. In turn, primary receptions of Alligator Alcatraz images shift to concerns about whether AI was used to doctor images or, in the case of swamp slop, the (in)appropriateness of the humor. This flattens the impact of, to keep with our example above, reports of a moderate amount of water leaking into the facility. It also flattens the scale and temporal dimensions of harm. An unspectacular amount of reported flooding does not preclude the levels of harm illustrated by the flood deepfake. AI changes the standard of what elicits sustained public outcry in the expansion of the detention and deportation regime.  

---

Image Credits:

1. Author’s screenshot of an AI Image of alligators in ICE hats from Offical DHS X Account.
2. Author’s screenshot of an X post from Tony Lane containing an AI image of the facility with an alligator-filled moat surrounding it.
3. Author’s screenshot of an X post of an AI Video showing the flooding of the ICE facility.
4. Author’s screenshot of X post from Jason Delgado about the actual event of water seeping into the facility which chat bots have been confusing with the AI flood video.