Current Affairs
AI-Generated and Fabricated Media Spreading as Legitimate News
The boundary between authentic footage and fabricated media has never been harder to locate. Deepfakes and misinformation now circulate at a pace that outs

The boundary between authentic footage and fabricated media has never been harder to locate. Deepfakes and misinformation now circulate at a pace that outstrips most newsrooms’ ability to respond, with AI-generated images, audio clips, and synthetic video segments spreading across social media platforms before fact-checkers can even identify the source material. What was once a theoretical concern about the integrity of the visual record has become a daily operational problem for journalists, researchers, and ordinary news consumers alike.
How Generative AI Became a Misinformation Engine
The mechanics behind today’s AI-generated misinformation are worth understanding clearly, because the tools involved are not exotic or difficult to access. Generative AI models — the same category of technology that powers widely available image and text generators — can now produce photorealistic stills, convincing video clips, and synthetic voice audio in minutes. What once required a professional visual effects studio can now be accomplished with a consumer-grade laptop and a free or low-cost subscription to any number of platforms.
This accessibility has had measurable consequences. Research published via the National Institutes of Health’s PubMed Central identifies AI-enabled generative models and engagement-optimization algorithms as playing a central role in both the production and the amplification of disinformation — a pairing that is especially damaging. A fabricated image or video does not need to be technically perfect to spread; it only needs to be plausible enough to provoke a share before anyone looks closely. Social platform algorithms, designed to surface content that generates strong reactions, are structurally well-suited to accelerating exactly that dynamic.
The scale of the infrastructure supporting this phenomenon is significant. NewsGuard, a media-monitoring organization that tracks AI-generated content farms, has identified nearly 3,750 such sites operating across the web — automated or semi-automated operations that publish AI-generated text, images, and sometimes video at volume, mimicking the appearance of legitimate news outlets. These sites function as laundering mechanisms: content originating from a clearly synthetic source gains a veneer of credibility once it appears on a site that looks, at a glance, like a regional news outlet.
The Visual Record Under Pressure
One of the most consequential effects of the current wave of AI-generated media is its impact on the visual record of current events. Photographs and video have long served as primary evidence in journalism, in legal proceedings, and in historical documentation. When fabricated imagery circulates alongside genuine footage — and when the two are visually indistinguishable to the untrained eye — the evidentiary value of visual media is degraded across the board.
→ Read also: AI-Generated Images and Deepfakes Circulating as News
This is not a hypothetical concern. Rutgers University Libraries have documented a marked increase in AI-generated content contributing to misinformation online, and note specifically that recent news events — including coverage of the Iran conflict — have seen surges of AI-generated video circulating as though it were authentic battlefield or protest footage. The problem is compounded by the speed of news cycles: a synthetic video shared in the first hours after a breaking event can shape public understanding before any correction reaches the same audience.
Hyper-realistic AI-generated news videos are now flooding social media platforms in sufficient volume that even experienced media consumers report difficulty distinguishing real reporting from fabricated segments. The mimicry extends beyond visuals: synthetic audio, generated to replicate the cadence and vocal qualities of real broadcasters or public figures, is increasingly used to create fake news segments that carry the surface markers of institutional credibility — a recognizable news chyron, a familiar broadcast format, a voice that sounds authoritative.
Public Figures as Targets
Public figures occupy a particularly exposed position in this landscape. Their faces, voices, and public statements are extensively documented in training data, which makes it technically straightforward to generate synthetic media that places them in fabricated scenarios. The damage such content can cause is asymmetric: a fabricated video or image can be created in minutes and shared to millions of people within hours, while corrections — even when issued promptly by credible sources — rarely reach the same audience at the same speed.
The phenomenon is not limited to political figures. In May of the current period, Katy Perry posted an AI-generated image to her Instagram account, an incident that illustrated how even celebrities with large, engaged audiences can inadvertently contribute to the spread of synthetic media when the origin of an image is unclear. The episode was widely cited as an example of how AI-generated imagery has become sufficiently normalized that it circulates without immediate suspicion even in high-visibility contexts.
When fabricated content targets political figures specifically, the potential for real-world harm extends beyond reputational damage. Synthetic media that falsely depicts a political leader being attacked, making inflammatory statements, or engaging in illegal conduct can affect public opinion, influence electoral behavior, and in some cases provoke physical responses. The challenge for fact-checkers is that deepfakes and misinformation of this kind are often timed to coincide with genuine news events, making the fabrication harder to identify because it arrives in a context where unusual developments are already expected.
Detection, Literacy, and the Limits of Both
The response to AI-generated misinformation has developed along two parallel tracks: technical detection and media literacy education. Both are necessary; neither is sufficient on its own.
→ Read also: AI-Generated and Manipulated Images Circulate as Authentic Election and Political Content
Technical detection tools attempt to identify the artifacts left by generative AI — inconsistencies in lighting, unnatural skin texture, irregular blinking patterns in video, or the subtle statistical signatures that AI image generators leave in pixel distributions. Some platforms have begun embedding content provenance metadata in images and video, allowing downstream viewers to verify where a piece of media originated. These approaches are valuable but face a persistent limitation: as detection tools improve, so do the generative models they are designed to catch. The gap between generation capability and detection capability has, in practice, consistently favored the generators.
Media literacy programs, offered by institutions ranging from university libraries to public broadcasting organizations, focus on equipping news consumers with practical habits: checking the source before sharing, reverse-image searching photographs, looking for the telltale signs of AI generation, and treating viral content with proportional skepticism. Resources developed by institutions such as SFCC Library and Rutgers University Libraries provide structured guidance on spotting AI-generated content, covering visual cues, metadata examination, and source verification workflows. These programs are genuinely useful, but they place a significant cognitive burden on individual users and assume a level of engagement that most casual social media users do not apply to content they encounter in their feeds.
The Institutional Response and Its Gaps
Platforms, governments, and news organizations have each taken steps to address the problem, with varying degrees of effectiveness. Social media platforms have introduced labeling systems for AI-generated content, though compliance depends on creators voluntarily disclosing the synthetic origin of their material — a condition that is, by definition, unlikely to be met by those using AI to deceive. Regulatory frameworks in several jurisdictions now require disclosure of AI-generated content in political advertising, but enforcement mechanisms remain underdeveloped.
News organizations have responded by strengthening verification protocols and investing in provenance tools, but the economics of digital media create structural pressures that work against thoroughness. Speed is rewarded; caution is not always compatible with being first. The result is that the burden of verification often falls on the audience rather than on the institutions nominally responsible for filtering the information environment.
Research published through the Centre for Economic Policy Research suggests that trusted news sources retain measurable value precisely because they provide a credible counter-signal in an environment saturated with synthetic and manipulated content. The implication is that institutional credibility — built slowly through consistent accuracy — represents one of the most durable defenses against the effects of AI-generated misinformation, even as that credibility is itself a target for fabricated media designed to undermine it.
Understanding deepfakes and misinformation as a structural feature of the current information environment, rather than a series of isolated incidents, is the necessary starting point for any serious response. The specific cases that surface each week — fabricated protest footage, synthetic news segments, AI-generated images of public figures — are symptoms of a deeper shift in what it costs to produce convincing false media and what it takes to correct the record once that media has spread. Addressing the symptom without understanding the mechanism produces responses that are perpetually one step behind.
This article was produced with AI assistance and reviewed editorially.
