Familiarity breeds complacency, not competence
In Brazil, one of the world's most digitally engaged societies, a troubling paradox has emerged: widespread immersion in AI-generated video has not produced the ability to recognize it. Researchers have found that despite constant exposure to synthetic media, Brazilians cannot reliably distinguish deepfakes from reality — a gap that carries serious consequences for a nation already navigating fragile information ecosystems and contested democratic processes. Familiarity, it turns out, is not the same as understanding, and comfort with synthetic content may be quietly eroding the very skepticism needed to resist it.
- Researchers have confirmed what many feared: high exposure to deepfakes does not translate into the ability to detect them, and Brazil's digitally saturated population is dangerously underprepared.
- The tools to create convincing synthetic video are now accessible to nearly anyone, while the public's capacity to question what it sees has not kept pace — a widening asymmetry with real consequences.
- Brazil's fragmented media landscape, where WhatsApp and social platforms carry more daily news than traditional outlets, creates ideal conditions for a single convincing deepfake to outrun any correction.
- People believe their intuition has been sharpened by exposure, but that confidence is itself a vulnerability — especially when synthetic content confirms what a viewer already wants to believe.
- Fact-checkers, platform detection tools, and media literacy campaigns all exist, but none can yet match the speed or scale at which sophisticated synthetic media is being produced and spread.
Brazil presents a paradox that, on closer inspection, reveals something genuinely alarming. Brazilians are among the world's most active consumers of online video — scrolling through reels, clips, and shorts, vast quantities of which are AI-generated. Yet when researchers tested their ability to distinguish real footage from deepfakes, the results were sobering. Exposure had not produced competence. People had grown comfortable inside a synthetic media ecosystem without developing the literacy to recognize when they were being deceived.
This gap sits at the intersection of two accelerating forces. Deepfake technology has become dramatically more accessible — no studio, no specialized training required. At the same time, AI-generated content has been normalized into the ambient texture of daily digital life. The intuition people believe they have developed turns out to be largely illusory. A convincing synthetic video of a political figure, a business leader, or a celebrity can be crafted well enough that even a practiced eye misses the seams — and if the content aligns with a viewer's existing beliefs, the motivation to question it shrinks further.
The stakes are especially high in Brazil, which has already weathered significant waves of election-related misinformation and where social platforms like WhatsApp, TikTok, and Instagram function as primary news sources for large portions of the population. The infrastructure for spreading synthetic media is fully in place. What is missing is the collective capacity to resist it.
No clear solution is on the horizon. Media literacy campaigns have shown mixed results globally. Fact-checking organizations cannot scale to the volume of content being generated. Platform detection tools perpetually lag behind the technology they are meant to catch. What the research makes plain is that exposure alone will never close this gap. What Brazilians need — and what remains largely absent — is a deeper understanding of how synthetic media is made, how it exploits perception, and why skepticism toward video evidence is not paranoia but necessity.
Brazil has a problem that looks like a paradox on the surface but reveals something darker underneath. Brazilians spend enormous amounts of time online, scrolling through videos, reels, and clips—many of them synthetic, many of them created by artificial intelligence. Yet when researchers tested whether people could actually tell the difference between a real video and a deepfake, the results were sobering. The ability to detect these manipulated videos was far worse than exposure alone would suggest. People who lived inside the ecosystem of synthetic media could not reliably identify it.
The disconnect matters because it sits at the intersection of two accelerating trends. On one side, the technology for creating convincing deepfakes has become dramatically more accessible. You no longer need a studio or specialized equipment or years of training. On the other side, the Brazilian public has become saturated with AI-generated content—it's woven into the daily scroll, normalized, ambient. You would think that familiarity would breed competence. Instead, the research suggests the opposite: people have become comfortable with synthetic video without developing the literacy to recognize when they're being deceived.
This gap between exposure and detection ability is not merely an academic concern. Brazil has experienced significant waves of election-related misinformation in recent years. Deepfakes have already been weaponized in political campaigns elsewhere in the world. The country's media landscape is fragmented, with significant portions of the population relying on messaging apps and social networks for news rather than traditional outlets. In that environment, a convincing synthetic video can spread faster than any correction, and a population that cannot reliably identify manipulation becomes vulnerable to it.
The research reveals a troubling asymmetry. People believe they can spot a fake. They have been exposed to so much synthetic content that they feel they have developed an intuition for it. But intuition and actual ability are not the same thing. A deepfake of a political figure saying something inflammatory, or a business leader making a damaging admission, or a celebrity in a compromising situation—these can be crafted with enough sophistication that the human eye, even the practiced eye, cannot catch the seams. And if the viewer already believes the content aligns with their expectations or their politics, the cognitive work required to question it becomes even harder.
What makes this particularly acute in Brazil is the scale and speed of digital adoption. The country has one of the world's largest and most engaged social media populations. TikTok, Instagram, YouTube, WhatsApp—these are not peripheral to Brazilian life. They are central. The infrastructure for spreading synthetic media is already in place. The audience is already there. What is missing is the collective ability to resist it.
The path forward is not obvious. Media literacy campaigns have shown mixed results globally. Fact-checking organizations exist but cannot scale to the volume of content being created. Platform companies have invested in detection tools, but those tools are always one step behind the technology that creates the fakes. What seems clear is that exposure alone will not solve the problem. Brazilians need not just to see more synthetic videos, but to understand how they are made, what techniques are used to fool the eye, and how to approach video evidence with appropriate skepticism. Without that understanding, the sophistication of the technology will only widen the gap between what people think they can detect and what they actually can.
The Hearth Conversation Another angle on the story
Why does exposure to deepfakes not make people better at spotting them? Shouldn't practice improve the skill?
You'd think so, but it seems like familiarity breeds complacency instead. When something is everywhere, you stop questioning it. You assume you know what you're looking at.
So the problem is psychological, not technical?
Partly. But it's also about the speed of the technology. Deepfakes are getting better faster than people's intuition is developing. You're always playing catch-up.
What happens if someone shares a convincing deepfake of a politician right before an election?
It spreads before anyone can verify it. In Brazil's media landscape, where messaging apps are how people get news, a single video can reach millions before any fact-checker even sees it.
Is there a way to fix this?
Media literacy helps, but it's slow work. The real issue is that the technology is outpacing our ability to regulate or understand it. We're in a moment where the tools exist to deceive, but the defenses don't.
So what should people do right now?
Be skeptical of anything that confirms what you already believe. That's usually where deepfakes do their most damage—not by showing you something impossible, but by showing you something you wanted to believe was true.