How Generative AI, State Communication, and the Attention Economy Are Changing Political Reality

 

Political deepfakes are only one part of a much larger transformation in digital political communication. Generative AI, rage bait, meme culture, engagement farming, and state-sponsored shitposting increasingly operate within the same information environment. The central problem is therefore not simply whether a fabricated image can fool someone. It is whether citizens can still determine what evidence deserves trust when political communication itself rewards provocation, ambiguity, and emotional manipulation.

 

The consequences extend beyond misinformation. Synthetic media can obscure authentic eyewitness evidence, innocent people can be falsely identified, governments can communicate through intentionally provocative memes, and audiences can become increasingly uncertain about whether any image or video should be believed.

 

At the same time, deepfakes should not automatically be treated as uniquely persuasive or inherently illegitimate. Empirical research challenges some of the strongest claims about their persuasive power, while political satire remains an important form of democratic expression. The ethical problem therefore requires distinctions among deception, parody, propaganda, political power, and legitimate criticism.

 

This article examines five questions:

• What distinguishes rage bait, shitposting, and political deepfakes?
• Why can synthetic media damage public knowledge even when people do not believe individual deepfakes?
• What happens when governments adopt the communication techniques of internet trolls?
• When does political satire serve democratic criticism, and when does it become deceptive manipulation?
• What ethical and policy problems remain unresolved?

 

What Is Rage Bait, and Why Has It Become Politically Important?

Rage bait is online content deliberately structured to provoke anger or frustration because emotional reactions generate attention, comments, sharing, and traffic. Oxford University Press selected “rage bait” as its 2025 Word of the Year, reflecting a broader transition from curiosity-based clickbait toward content engineered around emotional activation (Silva, 2025). The distinction matters because the economic incentive has changed. Traditional clickbait asks: Can I make you curious enough to click? Rage bait asks: Can I make you angry enough to react?

 

The second strategy can be considerably more corrosive to political discourse because the anger itself becomes part of the distribution mechanism. A person does not have to agree with a post to make it successful. Condemning it, mocking it, reposting it, or arguing beneath it can all increase its circulation. This creates what can be called an outrage loop. A provocative claim generates responses, those responses become evidence that a controversy exists, and the controversy then produces additional content about the reactions to the original content. Political communication consequently becomes less about informing audiences and more about maintaining emotional engagement (Daily Targum Editorial Board, 2025). The underlying incentive is structurally important. A content producer may benefit whether an audience responds with approval or disgust.

What Is Political Shitposting?

Political shitposting is deliberately provocative, frequently low-context digital content designed to entertain an in-group while confusing, irritating, or baiting an out-group.

What began largely as an internet subcultural practice has increasingly entered institutional political communication (Topinka, 2025).

Political shitposting is not identical to misinformation. A successful shitpost may be obviously absurd. Its political function can instead come from its ability to collapse the distinction between argument and entertainment. This creates a rhetorical asymmetry.

Suppose a government agency publishes an inflammatory meme about a serious policy matter. A critic has several options. The critic can ignore it, allowing the framing to circulate uncontested. The critic can respond seriously, which amplifies the original provocation. Or the critic can participate in the joke, thereby accepting the communicative rules established by the original poster. In each case, the shitposter controls the frame. That is why shitposting can be politically effective without offering a substantive argument.

 

How Has Shitposting Entered Official Government Communication?

The boundary between internet trolling and official political communication has become increasingly difficult to maintain.                                      The research documents government social-media posts that transform immigration enforcement, political opposition, military action, and presidential symbolism into entertainment formats normally associated with meme culture.

Examples include:

• Department of Homeland Security footage of immigration arrests edited to the Pokémon theme and the phrase “Gotta Catch ’Em All.”
• Immigration-enforcement footage paired with podcaster Theo Von’s remark, “Heard you got deported dude, bye.”
• A White House display replacing Joe Biden’s presidential portrait with an image of an autopen.
• An altered arrest photograph of civil-rights attorney Nekima Levy Armstrong in which artificial tears were added to her face.
• Political videos using popular music and entertainment references to provoke artists, critics, and political opponents.

These are not merely changes in graphic design. They represent a change in the communicative relationship between governmental institutions and citizens.

Government communication traditionally carries an epistemic expectation, meaning citizens reasonably expect official institutions to distinguish factual reporting from parody, advocacy, and fabrication. When an official account deliberately blurs those categories, the problem is not simply bad taste. It is an accountability problem.

Why Does State-Sponsored Shitposting Matter Ethically?

State-sponsored shitposting matters because the moral status of communication changes when the speaker possesses institutional power. Satire directed upward at powerful institutions performs a different social function from ridicule directed downward by those institutions against individuals subject to their authority. That distinction becomes especially important when the government controls policing, immigration enforcement, detention, prosecution, or military force.

Consider an altered photograph of a person who has been arrested. If an ordinary citizen creates an obvious parody, questions of satire and free expression immediately arise. If the government alters the photograph and distributes it through an official account while the individual remains subject to governmental authority, additional ethical concerns appear.

The relevant questions include:

 

• Is the alteration clearly disclosed?
• Is the subject a public official or a private citizen?
• Could the image create a false impression concerning an actual event?
• Is the government commenting on someone against whom it possesses enforcement or prosecutorial power?
• Does the communication interfere with public understanding of the underlying facts?
• Can officials avoid accountability merely by describing realistic manipulation as a “meme”?

These questions cannot be resolved by saying that political satire has constitutional protection. The ethical issue concerns not only what expression is legally permissible, but also what communicative obligations attach to public power.

What Role Does Generative AI Play in Political Rage Bait?

Generative AI radically lowers the cost of producing emotionally provocative synthetic media. A creator no longer needs professional visual-effects skills to generate videos depicting immigration raids, political violence, military action, arrests, or public figures. Synthetic material can be produced rapidly, replicated across accounts, and optimized for emotionally charged distribution.

404 Media documented Sora-generated videos depicting fictional immigration raids that accumulated enormous engagement on Facebook. Some portrayed workers being detained or mothers being separated from infants. Visual defects were sometimes apparent, yet the videos still circulated widely (Koebler, 2025).

 

An important assumption should be challenged here. The existence of convincing synthetic media does not establish that sophisticated deepfakes are uniquely persuasive. Michael Hameleers’ research, as summarized in the supplied research materials, found that political deepfakes were not necessarily judged as more credible than textual misinformation and could be perceived as less credible than simpler manipulated media. That finding complicates the common narrative that increasingly realistic AI automatically produces increasingly effective deception. The political danger may lie elsewhere.

Are Deepfakes Really More Persuasive Than Ordinary Misinformation?

Not necessarily. Research summarized in the underlying report indicates that people may approach explicitly AI-like political media with considerable skepticism. Deepfakes can contain visual inconsistencies, mismatched voices, implausible behavior, or other cues that trigger suspicion.

 

Simpler manipulation can sometimes be more effective. A cheapfake does not require generative AI. It may involve selectively editing authentic footage, changing playback speed, removing context, using misleading captions, or combining genuine material in deceptive ways. Such content possesses an important advantage: much of what the audience sees is actually real.

 

The ethical debate about synthetic media therefore risks becoming distorted if policymakers focus narrowly on technologically sophisticated deepfakes while ignoring familiar techniques of misrepresentation. The relevant distinction is not AI versus non-AI. The relevant distinction is Reliable representation versus deceptive representation.

How Can Deepfakes Damage Truth Even When People Do Not Believe Them?

Synthetic media can damage an information environment without successfully deceiving most viewers. This is one of the most important distinctions in the deepfake debate. The harm can occur through at least four mechanisms:

1. Algorithmic dilution of genuine evidence
2. The liar’s dividend
3. False identification through AI enhancement
4. Epistemic exhaustion and news avoidance

Each mechanism concerns the broader information environment rather than the persuasiveness of one fabricated image.

How Can AI Content Drown Out Real Eyewitness Evidence?

Authentic evidence now competes with synthetic content for the same finite resource: human attention. Social-media platforms reward material capable of producing large numbers of views, shares, comments, and emotional reactions. Sensational synthetic videos may therefore receive more circulation than authentic but less spectacular recordings.

The problem is algorithmic dilution. When hundreds or thousands of fabricated images surround a real event, authentic evidence becomes one more item inside an information torrent. A citizen attempting to determine what happened must distinguish:

• genuine eyewitness recordings,
• AI-generated scenes,
• edited authentic footage,
• satire,
• partisan commentary,

 

• recycled footage from unrelated events,
• and fact-checks responding to each

The informational burden has shifted from the producer to the viewer. That burden is not trivial.

What Is the Liar’s Dividend?

The liar’s dividend occurs when widespread awareness of synthetic media allows genuine evidence to be dismissed as fabricated. Deepfakes therefore produce two opposite epistemic risks. The first is believing something false because fabricated media appears authentic. The second is rejecting something true because authentic media can plausibly be called fake.

The second danger is easy to underestimate. Once citizens know that realistic images, voices, and videos can be fabricated, a politician, corporation, government agency, or private citizen confronted with damaging authentic evidence gains a ready response, “That is AI-generated.” The mere existence of deepfake technology can therefore provide rhetorical protection to liars even when no deepfake has actually been used.

The political problem is deniability, not deception alone.

Why Is AI “Enhancement” Particularly Dangerous?

 

Generative AI does not recover visual information that was never captured. This distinction is essential when AI tools are used to “enhance,” “unmask,” or reconstruct faces from low-quality images. Following the fatal shooting of Renee Good, social-media users reportedly used xAI’s Grok in attempts to reconstruct the face of a masked federal agent. The resulting synthetic imagery contributed to false identification and harassment of innocent people sharing the suspected individual’s name (Brumfiel, 2026).

An AI system can produce a visually plausible face without producing the correct face. Greater visual clarity is therefore not necessarily greater evidentiary accuracy. This is a fundamental media-literacy principle: Synthetic detail should not be confused with recovered evidence. A blurry photograph honestly represents the limits of the available information. A photorealistic AI reconstruction may conceal those limits.

What Happens When Synthetic Political Trauma Becomes a Business Model?

Some of the most disturbing synthetic political videos may not originate from ideological campaigns at all. Some highly provocative synthetic immigration videos appear to have been produced by overseas engagement farmers seeking advertising or platform revenue rather than political influence. This distinction matters. A video can

 

have major political effects even when its creator possesses no meaningful political objective. The causal sequence can instead be:

platform incentive → emotionally extreme content → audience reaction → algorithmic amplification → political interpretation.

The creator needs only to understand that American political conflict generates engagement. This means that disinformation analysis should not always begin by asking, “Who benefits politically?” Sometimes the more useful question is, “Who gets paid when people become angry?”

Can Political Opponents and Foreign Actors Exploit the Same Meme Ecosystem?

Yes, at least in the broader structural sense documented by the research materials.

Domestic political organizations, state institutions, foreign propaganda operations, influencers, and commercial engagement farmers all compete within platforms whose primary distribution currency is attention. They therefore face similar incentives. The result is convergence. Different actors can independently adopt:

• gaming imagery,
• children’s animation aesthetics,
• surreal AI videos,
• short-form humor,
• ironic language,
• celebrity references,
• deliberately inflammatory captions,
• and re-mixable meme

Ideological coordination is not required. An effective meme can be copied by ideological supporters, political opponents, commercial engagement farmers, and foreign propaganda networks for entirely different reasons. Distribution itself becomes decentralized. A user does not need to believe propaganda to propagate it. The user only needs to find it sufficiently outrageous, entertaining, or bizarre to share.

Why Is Gamification So Effective in Political Propaganda?

Gamification removes some of the cognitive signals people normally associate with formal political persuasion. A government editorial looks like propaganda. A meme may look like entertainment. A military strike presented through a video-game interface, animated character, or pop-culture reference can therefore change the audience’s psychological relationship with the event. The underlying event may involve death, coercion, arrest, or war. The representational form communicates play.

 

This does not prove that audiences become unaware of the event’s seriousness. The existing evidence demonstrates a striking rhetorical transformation: governmental violence can be represented using the visual grammar of games, jokes, animation, and fandom. The ethical question is therefore unavoidable. “What happens to democratic accountability when state violence becomes content?”

Does Political Humor Always Undermine Democracy?

Obviously not. Political satire has a long history of challenging authority, exposing hypocrisy, and allowing people without institutional power to ridicule those who possess it. Research and advocacy describe synthetic political satire as a potential “weapon of the weak.” Deepfake parody can allow artists, activists, and dissidents to appropriate the image of a powerful political figure and undermine carefully controlled political symbolism.

 

This function is especially important in environments where governments dominate conventional communications channels. The ethical distinction therefore cannot simply be:

serious political speech = legitimate

and

synthetic political humor = illegitimate.

Context matters.

 

 

When Does Political Satire Become Ethically Problematic?

Political satire becomes more ethically difficult when the content is realistic enough to be mistaken for documentary evidence and when the person distributing it possesses substantially greater power than the person represented. Several variables matter:

• Disclosure: Is the manipulation apparent?
• Realism: Could a reasonable viewer mistake it for documentation?
• Target: Is the subject a powerful public official or an ordinary citizen?
• Purpose: Is the goal criticism, entertainment, humiliation, deception, or intimidation?
• Context: Does the synthetic content concern an actual arrest, prosecution, military event, or other high-stakes occurrence?
• Power: What institutional authority does the creator possess over the subject?
• Correctability: Can the false impression realistically be corrected after viral distribution?

 

These distinctions explain why a blanket prohibition on political deepfakes would be ethically crude. They also explain why labeling every manipulated government image “satire” is inadequate.

What Is Wrong With the “It’s Just a Meme” Defense?

Calling something a meme describes its cultural format. It does not resolve its truth status or ethical significance. A realistic altered photograph can simultaneously be:

• a meme,
• politically motivated,
• humorous to an intended audience,
• factually deceptive,
• and harmful to its

These categories are not mutually exclusive. The phrase “just a meme” can therefore function as an accountability shield. Critics object to the deception. Defenders respond by changing the classification of the communication from factual representation to humor. But the central question remains unanswered: Was the audience likely to understand what was real and what was fabricated? If not, the meme label does little ethical work.

How Does Rage Bait Damage Rational Political Discourse?

Rage bait rewards immediate emotional response over evidentiary evaluation. That incentive reverses the normal sequence of sound reasoning. Critical reasoning should begin by identifying facts, evaluating evidence, separating factual claims from value judgments, and only then forming a normative conclusion.

Rage bait encourages the opposite sequence:

1. Experience
2. Select an
3. Share or
4. Defend the
5. Investigate the facts later, if at

This is epistemically backward. It also makes logical fallacies commercially valuable. Loaded language, false dilemmas, hasty generalizations, ad hominem attacks, guilt by association, and selective evidence can all outperform careful argument when engagement rather than truth is the operative metric. The problem is the platform incentive can reward bad argumentation.

 

Can Rage Bait Target Political Allies as Well as Opponents?

Yes. Outrage does not require an ideological enemy. It requires a perceived violation. Political organizations can therefore provoke supporters by presenting members of their own coalition as traitors, cowards, sellouts, or insufficiently committed. Fundraising tactics directed at fellow conservatives is one example where perceived betrayal can itself become commercially useful.

This demonstrates an important feature of outrage economies. They do not merely polarize opposing groups. They can continually fragment groups internally by rewarding purity tests and betrayal narratives. A political ecosystem organized around anger eventually needs new objects of anger.

Can Synthetic Political Media Desensitize Audiences?

The research raises this question as a significant hypothesis, but the research does not provide longitudinal evidence sufficient to establish that repeated exposure to synthetic political violence causes durable reductions in empathy or political concern.

Repeated exposure to fabricated trauma could plausibly generate fatigue, habituation, skepticism, or avoidance. However, establishing those effects requires empirical evidence rather than intuition. The research identifies this as a major gap and calls for longitudinal studies examining whether repeated exposure to simulated violence changes empathy, willingness to intervene, or political support for victims.

Why Might Synthetic Media Increase News Avoidance?

An information environment can become unusable before it becomes completely deceptive. Consider what is required from an ordinary person encountering a viral political video. The viewer may have to determine:

• whether the event occurred,
• whether the footage is authentic,
• whether the audio belongs to the video,
• whether AI has modified any portion,
• whether the caption describes the event accurately,
• whether the account is credible,
• whether fact-checks exist,
• and whether those fact-checks are themselves

At some point, disengagement becomes rational from the individual’s perspective.

The person does not necessarily conclude that everything is false. The person concludes that determining what is true requires more time than they are willing or able to spend. This is an epistemic cost of synthetic media that simple true-versus-false models overlook.

 

What Are the Legal Risks of State-Sponsored Meme Politics?

Different legal issues arise depending on the content. Potential areas include:

• copyright infringement involving music or other protected media,
• unauthorized use of likeness or voice,
• defamation,
• false-light claims,
• right-of-publicity disputes,
• deceptive synthetic media statutes,
• and constitutional questions involving government treatment of individuals under investigation or prosecution.

Multiple instances exist in which artists and other public figures objected to government use of their work or likeness, followed in some cases by removal or alteration of social-media content. However, one especially important constitutional question remains unresolved in the supplied materials.

Could Government Deepfakes Create a Due-Process Problem?

My research does not establish whether an official government entity’s use of realistic deceptive synthetic media about a person under arrest violates Fifth Amendment due-process protections. The research explicitly identifies this as an empirical gap. The issue deserves serious legal examination.

Imagine that a federal agency arrests someone and then distributes a realistic altered image designed to make that person appear frightened, dishonest, unstable, or contemptible. Contemptability was the goal the case of adding artificila tears to the arrest photograph of civil-rights attorney Nekima Levy Armstrong. When the government creates such an environment, he government is no longer merely commenting on a public controversy. The government possesses coercive authority over the person depicted.

Questions then arise concerning prejudice, governmental neutrality, reputational injury, fair-trial interests, and the relationship between official speech and due process. A definitive constitutional conclusion would require legal authority not contained in the present research.

What Is the Most Important Mistake in the Deepfake Debate?

The most important mistake may be treating deepfakes primarily as fake videos that fool people. Such a view is too narrow. A more thorough analysis concerns the epistemic environment: the social system through which people decide what deserves belief. Generative AI changes the epistemic environment by making several things easier simultaneously:

 

• manufacturing fabricated evidence,
• manufacturing entertainment from serious events,
• creating multiple contradictory versions of reality,
• falsely “enhancing” poor-quality evidence,
• overwhelming authentic material with synthetic material,
• and dismissing authentic evidence as

The resulting problem is about the status of evidence itself.

What Ethical Principles Should Guide Political Synthetic Media?

A useful ethical framework should distinguish between the legitimacy of satire and the obligations created by deception and institutional power. Political synthetic media must be held to higher standards due to factors previously discussed (due process concerns, institutional power and coercion, etc.). At minimum, political synthetic media should be evaluated according to the following principles.

1.  Truthfulness

Does the communication create a materially false belief about a real person or event?

2.  Transparency

Can an ordinary viewer reasonably identify the synthetic or altered character of the content?

3.  Power

Does the creator possess institutional, political, economic, or coercive power over the person depicted?

4.  Harm

What foreseeable reputational, physical, legal, political, or social harms could follow from circulation?

5.  Public Interest

Does the manipulation contribute to legitimate criticism, artistic expression, or public understanding, or primarily distort factual understanding?

6.  Correctability

Can false impressions realistically be corrected after distribution?

 

7.  Evidentiary Integrity

Could the synthetic material interfere with the public’s ability to evaluate authentic evidence concerning the same event?

 

 

These questions avoid two overly simplistic conclusions: The first is that synthetic political media is inherently unethical. The second is that anything labeled parody or meme is ethically harmless. Neither position withstands serious analysis.

What Responsibilities Do Government Institutions Have?

Government institutions should be held to a higher standard of representational accuracy than ordinary meme creators when communicating about real exercises of governmental power. Government agencies possess authority that ordinary speakers do not possess. They can arrest, prosecute, deport, regulate, and use force. The epistemic consequences of deceptive official communication are therefore different from those of an anonymous parody account.

A workable governmental standard would include:

• clear labeling of synthetic or materially altered media,
• preservation of unedited source material,
• separation between documentary communication and obvious parody,
• restrictions on realistic manipulation of identifiable private citizens,
• transparent correction procedures,
• and archival preservation of deleted or modified official

Such standards would not eliminate political humor. These standards would establish boundaries between institutional communication and manufactured evidence.

What Responsibilities Do Social-Media Platforms Have?

Platforms occupy a structurally difficult position because they simultaneously host political expression, commercial advertising, entertainment, journalism, propaganda, and synthetic media. A responsible approach should therefore focus less on banning categories of speech and more on preserving informational context. Relevant measures include:

• provenance systems for authentic media,
• conspicuous disclosure of materially synthetic political content,
• preservation of original upload metadata where possible,
• friction before redistribution of disputed content,
• reduced financial incentives for engagement-farmed deceptive media,

 

• access for independent researchers studying recommendation systems,
• and rapid correction mechanisms when synthetic content falsely identifies private individuals.

The unresolved issue is incentives. If outrage remains profitable, moderation alone addresses symptoms while leaving the economic structure intact. Should be designed and created almost exclusively for emotional engagement be monetizied? A negative answer to this question would also include viral ASMR videos (videos designed to produce a calming effect that reduces stress, etc.) that are nonpolitical and not rage bait or shitposting.

What Should We Ultimately Worry About?

We are entering an environment in which genuine videos, cheapfakes, AI-generated images, satire, government propaganda, engagement farming, eyewitness recordings, political memes, and foreign information operations circulate through the same channels and compete according to similar attention incentives. Truth does not disappear in such an environment. Truthfulness becomes expensive to establish.

Democratic societies depend on citizens being able to criticize authorities, document abuses, parody powerful leaders, disagree about values, and inspect evidence supporting factual claims. Generative AI does not make those practices impossible. Gen AI makes the norms surrounding them more important.

Political deepfakes should be judged according to what they do, whom they target, how they represent reality, what power their creators possess, and whether they strengthen or weaken the public’s ability to distinguish evidence from manipulation. The central ethical challenge of political synthetic media is not that machines can now create convincing fictions. People were using Photoshop long before Gen AI to make deepfakes and political satire. Gen AI allows for these to be produce more quickly, efficiently, and by more people. Political deception is also not new.

The deeper problem is that our contemporary information economy increasingly rewards the deliberate destruction of the boundary between evidence and entertainment. When governments, political campaigns, propagandists, commercial engagement farmers, and ordinary users all benefit from making that boundary harder to see, media literacy alone cannot solve the problem. The institutions governing political communication must decide whether truth is merely another form of content or a public good worth protecting.

 

Frequently Asked Questions

What is a political deepfake?

A political deepfake is synthetic or substantially AI-manipulated media depicting political figures, events, or issues in ways that can alter how audiences interpret reality. Deepfakes may be deceptive, satirical, artistic, or propagandistic. Their ethical status depends heavily on context, disclosure, intent, target, realism, and potential harm.

What is rage bait?

Rage bait is digital content deliberately designed to provoke anger or frustration because emotional reactions increase attention, traffic, comments, and sharing. Unlike traditional clickbait, which primarily exploits curiosity, rage bait turns emotional hostility itself into an engagement mechanism (Silva, 2025).

What is political shitposting?

Political shitposting is provocative, often low-context or deliberately absurd online content used to entertain supporters, bait opponents, and disrupt conventional political discourse. It becomes especially consequential when political campaigns or government institutions adopt these techniques as official communication strategies (Topinka, 2025).

Are deepfakes more believable than ordinary misinformation?

Not necessarily. Research summarized in the underlying report indicates that sophisticated AI deepfakes may sometimes be viewed as less credible than simpler manipulations because audiences are increasingly alert to AI-generated media.

Deepfake risk therefore should not be assessed solely according to technological sophistication.

What is the liar’s dividend?

The liar’s dividend is the advantage gained when widespread knowledge of deepfakes allows genuine evidence to be dismissed as fabricated. As synthetic media becomes common, people confronted by authentic but damaging recordings can claim that the material was AI-generated, even when it was not.

Why are AI-enhanced images unreliable as forensic evidence?

Generative AI can create plausible visual detail that was absent from the original recording. An AI-generated reconstruction may therefore look clearer without being more accurate. Synthetic visual detail should not be treated as recovered biometric or forensic information.

 

Can deepfakes be legitimate political satire?

Yes. Synthetic media can be used for parody, artistic criticism, and political dissent. Ethical evaluation should distinguish obvious satire directed at powerful actors from realistic deceptive media targeting private individuals or people subject to governmental authority.

Why is state-sponsored shitposting different from ordinary internet trolling?

Government institutions possess coercive and epistemic authority that ordinary users do not possess. When official accounts manipulate images or use misleading memes concerning arrests, enforcement, war, public policy or legitimate political viewpoints, or other state actions, the communication can affect public understanding of governmental conduct and accountability.

What is the main ethical problem created by political deepfakes?

The deepest problem is not simply that people may believe false videos. Synthetic media can also make authentic evidence easier to dismiss, flood information channels with competing fabrications, encourage false identification, and increase the effort required to determine what actually occurred.

 

References

This bibliography provides full academic-grade APA-style (7th edition) citations for all 25 sources in this notebook. The citations are ordered alphabetically by the author’s last name or organizational creator, in a flat list without annotations or thematic groupings.

—

1. Associated Press. (2026, August 13). Pro-Iran groups deploy AI to troll Trump, influence the war narrative. *PBS NewsHour*. https://www.pbs.org/newshour/world/pro-iran-groups-deploy-ai-to-troll-trump-influence-the-war-narrative
2. Berger, E. (2026, August 9). Taylor Swift songs removed from Trump and White House’s social media posts. *The Guardian*. https://www.theguardian.com/us-news/2026/aug/09/taylor-swift-songs-removed-trump-white-house-posts
3. Brumfiel, G. (2026, January 8). AI images and internet rumors spread confusion about ICE agent involved in shooting. *NPR / Texas Public Radio*. https://www.tpr.org/2026-01-08/ai-images-and-internet-rumors-spread-confusion-about-ice-agent-involved-in-shooting
4. Daily Targum Editorial Board. (2025, December 7). Editorial: Rage bait replaces discourse. *The Daily Targum*.
5. DiResta, R. (2026, April 2). When virality is the message: The new age of AI propaganda. *TIME*.
6. Gemini Notebook. (2026). *The computational ecology of anti-ICE synthetic media: Forensic analysis of generative AI, algorithmic monetization, and the geopolitical dimensions of strategic disinformation* [Markdown analysis].
7. Gemini Notebook. (2026, August 14). *Research report: Alternative and contrarian perspectives on political deepfakes* [Markdown database].
8. Gonzalez, E. (2026, May 7). Enraging encounters: Ragebaiting through public recordings ignores the respect and decency for others. *Branham Bear Witness*.
9. Gregory, S., & Cizek, K. (2021, December). *Just joking! Deepfakes, satire, and the politics of synthetic media*. MIT Open Documentary Lab’s Co-Creation Studio & WITNESS. https://cocreationstudio.mit.edu/wp-content/uploads/2021/12/JustJoking.pdf
10. Hameleers, M. (2024). Cheap versus deep manipulation: The effects of cheapfakes versus deepfakes in a political setting. *International Journal of Public Opinion Research*, *36*(1), 1–9. https://doi.org/10.1093/ijpor/edae004
11. Huamani, K. (2026). Trump’s use of AI images pushes new boundaries, further eroding public trust. *Associated Press*. https://apnews.com/article/ai-videos-trump-ice-artificial-intelligence-08d91fa44f3146ec1f8ee4d213cdad31
12. Koebler, J. (2025, November 12). Tech companies allegedly profiting from the dehumanisation of immigrant communities. *404 Media / Business & Human Rights Resource Centre*. https://www.business-humanrights.org/en/latest-news/meta-allegedly-profiting-from-the-dehumanisation-of-immigrant-communities/
13. Koebler, J. (2026, July 31). AI-Generated Sora Videos of ICE Raids Are Wildly Viral on Facebook. *404 Media*. https://www.404media.co/ai-generated-sora-videos-of-ice-raids-are-wildly-viral-on-facebook/
14. Mashable. (2026, January 22). White House uses AI to alter image of arrested anti-ICE protestor. *Mashable*. https://mashable.com/article/white-house-posts-ai-altered-image-minneapolis-church-protester
15. Metro Police Department. (2026). *Metro police warn of AI-generated videos amid anti-ICE protests in Southern Nevada* [Video]. YouTube.
16. MPR News. (2026). Trump’s use of AI images pushes new boundaries, further eroding public trust, experts say. *MPR News / Associated Press*. https://www.mprnews.org/story/2026/01/trumps-use-of-ai-images-pushes-new-boundaries
17. Myers, C. (2025, November 14). These viral ICE videos are actually AI — and that’s not even the most disturbing part. *HuffPost*. https://www.huffpost.com/entry/viral-ai-ice-raid-videos_l_69163baee4b0df3e16be4163
18. OpenAI. (2025, December 12). OpenAI responds to concerns about Sora’s alleged generation of dehumanising ICE raid videos. *Business & Human Rights Centre*. https://www.business-humanrights.org/en/latest-news/openai-responds-to-concerns-about-soras-alleged-generation-of-dehumanising-ice-raid-videos/
19. Peters, C. S., McGinnis, A. J., Im, S., & Hennessey, H. L. (2025). Satire versus illicit: Policy issues regarding the creation of deepfakes using AI technology. *Behavioral Sciences & the Law*, *43*(4), 462–471. https://doi.org/10.1002/bsl.2726
20. Silva, C. (2025, December 1). ‘Rage bait’ named word of the year. *Mashable*. https://mashable.com/article/rage-bait-word-of-the-year-2025
21. Steffen, S. (2026, February 1). Fact check: Are ICE fakes trying to drown out real videos? *Deutsche Welle / Inquirer.net*. https://globalnation.inquirer.net/307632/fact-check-are-ice-fakes-trying-to-drown-out-real-videos
22. Steffen, S. (2026, February 1). *Fact check: Are ICE fakes trying to drown out real videos?* [PDF Document]. Deutsche Welle.
23. Topinka, R. (2025, October 2). Do you know about shitposting? It’s cheap humour, rage bait – and now, it seems, US government policy. *The Guardian*. https://www.theguardian.com/commentisfree/2025/oct/02/shitpost-us-white-house-trump-joke
24. WSLS 10. (2026, February). Virginia Tech expert breaks down political motives of partial government shutdown. *WSLS 10*. https://www.wsls.com/news/local/2026/02/expert-federal-government-shutdown-DHS-ICE-politics
25. YouGov. (2025, October). Most Americans react negatively when shown three AI-generated videos posted by President Trump. *YouGov*.