502. How Short-Form Video Is Changing Psyops, Propaganda, and Online Influence

Email: bidemiologunde@gmail.com

In this episode, host Bidemi Ologunde examines how short-form clips from podcasts, songs, and movies are reshaping persuasion online. Why do thirty-second clips often feel more convincing than full arguments? How can entertainment, fandom, and creator culture become delivery systems for influence operations? What happens when AI-generated voices, synthetic video, and algorithmic feeds make propaganda faster, cheaper, and more emotionally targeted? Bidemi explores the healthy uses of clip culture, the risks of context collapse, and the emerging future of media literacy in an age where attention itself has become strategic terrain.

Transcript

SPEAKER_00 0:16

Forty-eight hours before Slovakia's twenty twenty three parliamentary election, an audio recording began moving through the country's online ecosystem. The recording appeared to capture Mikal Simeka, leader of the Liberal Progressive Slovakia Party, discussing election manipulation with journalist Monika Todova. The clip arrived during a vulnerable window when voters were already anxious, when institutions were already distrusted, and when fact checkers had very little time to respond. Later analysis complicated the claim that one deep fake swung the election, but the episode still gave the world a clear warning. Synthetic media does not need to convince everyone, and it does not even need to survive inspection forever. It only needs to move quickly enough, emotionally enough, and plausibly enough to shape the atmosphere before truth can catch up. A few months earlier, a different kind of synthetic clip had taken over the internet. In April 2023, a song called Heart on My Sleeve appeared on TikTok and streaming platforms, apparently featuring Drake and The Weeknd. The vocal sounded familiar, the hook worked, and millions of people encountered it as entertainment before they processed it as a legal and ethical problem. The artist had not made the song, and Universal Music Group pushed for its removal after it had already reached hundreds of thousands of Spotify streams, millions of TikTok views, and hundreds of thousands of YouTube views. That incident was not a military operation and it was not framed as political propaganda. It was still a preview of the same machinery, synthetic intimacy, platform acceleration, cultural familiarity, and a public that now has to ask whether a voice that feels real should be trusted. So those two stories set up today's big picture. Short form video has become the most efficient emotional delivery system in the modern media environment. Podcast clips, song fragments, movie scenes, influenza monologues, political edits, and synthetic videos now compete inside the same vertical feed. The format compresses context, rewards reaction, and turns cultural material into portable units of persuasion. When that system is healthy, it helps people discover ideas, art, communities, and long-form conversations. When that system is exploited, it becomes an unusually powerful surface for psychological manipulation. Let us define the problem carefully. Psychological oppressions or psyops are usually associated with military or state activity designed to influence emotions, attitudes, beliefs, and behavior. In everyday online culture, people use the word psy-up much more loosely, sometimes to describe any coordinated attempt to shape perception. For this episode, the useful framing is defensive and analytical. I'm asking how the rise of short clips could make influence operations more scalable, more believable, and harder to detect. The first major shift is that clips have become the new unit of attention. A long podcast episode may last two hours. A movie may last two hours, and a song may last three minutes. But the clip is what travels. A single sentence from a podcast can become a political identity marker. A chorus can become a mood. A movie scene can become evidence for a worldview it was never meant to support. The clip is portable, searchable, remixable, and emotionally dense. It moves well because it does not ask the audience for much time. It asks for a glance, a feeling, and a share. That matters because persuasion often begins before argument. A person does not always change their mind after hearing a formal case. Sometimes a person changes their mind after repeatedly encountering the same emotional frame from different sources. One clip makes an issue feel absurd, another makes it feel corrupt. Another makes it feel dangerous. Yet another makes one side feel confident and the other side feel pathetic. Over time, the viewer may not remember the exact source, but they remember the emotional temperature. That is where clip culture becomes important for influence operations. It can literally manufacture atmosphere. Podcast clips are especially useful in this environment because they feel intimate. A podcast conversation looks casual, unscripted, and human. The host leans back, the guest interrupts, someone laughs, and the audience feels as if it has overheard something more authentic than a prepared speech. This is part of the appeal of podcasting. It is also part of the risk. A 30-second clip can extract a claim from a two-hour conversation and detach it from the caveats, corrections, or uncertainty that surrounded it. Viewers meet the emotional punchline before they meet the full discussion. The scale of video podcasting helps explain why this matters right now. YouTube announced in February 2025 that more than 1 billion monthly active viewers were watching podcast content on the platform. Edison Research Infinite Dial 2025 found that 48% of Americans aged 12 and older had both listened to and watched a podcast, and that YouTube was the service used most often by US weekly podcast listeners. The podcast has become visual, searchable, algorithmic, and easy to cut into fragments. Songs operate differently, but they are just as powerful. A song clip does not usually persuade through argument, it persuades through repetition, identity, and mood. A song becomes a template, a lyric becomes a caption, a beat becomes the background for thousands of unrelated videos. If a movement, brand, faction, or government can attach itself to a sound, it can travel without looking like a message. People repeat it because it feels fun, stylish, nostalgic, rebellious, or communal. Music can turn persuasion into participation. Movie clips add another layer because they bring narrative authority. A scene from a film or television show can be used as shorthand for betrayal, heroism, corruption, masculinity, victimhood, revenge or social decay. The audience already understands the emotional grammar. When a clip from a movie is repurposed into a political or cultural argument, it imputs the emotional force of the original story. A character's rage can become the viewer's rage. A villain's monologue can become a meme. A fictional moment can become a way of interpreting real-life events. The entertainment industry has already noticed how powerful these fragments are. UGOV research, published in March 2026, found that short form video is becoming a discovery engine for television and film. Among people who had seen clips from shows or films on social media, 77% said they had gone on to watch the full program. Among 16 to 24-year-olds, 87% said they had started watching a full show or film after seeing clips or memes on social media. That is a marketing insight, but it is also an influence insight. Clips can move people from fragment to full commitment. The same logic applies to news and politics. Pure Research Center found in 2025 that 1 in 5 US adults regularly got news on TikTok, up from 3% in 2020. Among adults under age 30, 43% said they regularly got news on TikTok. That does not mean TikTok is only a news platform, and it does not mean that younger audiences are uniquely gullible. It means the boundary between entertainment feed and newsfeed has become porous. A person may encounter a comedian, a work clip, a beauty tutorial, a campaign message, and a disaster rumor in the same session. The short form environment has five features that make it attractive to influence operators. The first is speed. A clip can be produced, translated, captioned, remixed, and posted faster than institutions can explain complicated reality. Speed favors emotional certainty over careful verification. The second is volume. An operator does not need one perfect message when thousands of variations can be tested. The third is segmentation. Different audiences can receive different emotional versions of the same narrative. The fourth is ambiguity. A clip can be framed as humor, fan culture, criticism, education, commentary, or satire. The fifth is deniability. A manipulative message can be laundered through ordinary users, influencers, remixers, and fandom accounts. So this does not mean that every viral clip is suspicious. Most clips are ordinary culture moving through ordinary people. That distinction matters. The risk is not that users share things. The risk is that influence campaigns can hide inside normal sharing behavior. They can use real creators, real emotions, and real grievances as cover. A campaign does not always need to invent a sentiment from nothing. It can amplify a sentiment that already exists, sharpen it, and aim it. This is why the phrase context collapse is so important. Context collapse happens when a piece of content moves into an audience that does not share the assumptions of the original setting. A podcast clip meant for a niche audience enters a hostile political feed. A movie scene intended as satire becomes a sincere ideological symbol. A song meant as parody becomes a real anthem. A joke becomes evidence. A fragment becomes a worldview. The Letter to America incident in November 2023 showed how quickly a politically charged text could be rediscovered through short videos stripped of historical and moral context and turned into a platform controversy. Box described how videos featuring Osama bin Laden's 2002 letter were posted to TikTok, triggering condemnation from politicians, 9-11 families, and influencers while also noting that simplified narratives about the episode did not capture the whole reality. The incident was valuable as a case study because it showed the feedback loop between users, platforms, politicians, journalists, and moral panic. A platform trend became a national argument about radicalization, censorship, youth culture, and geopolitical perception. The deeper lesson is that short clips can create sudden interpretive communities. A person watches one video and then sees dozens of people reacting to the same object. The object becomes a shared test of identity. What do you think about this letter, this speech, this work clip, this police video, this celebrity comment, this movie scene, this podcast argument, and so on. The feed does not merely show content. It asks the viewer to declare allegiance through attention, comment, and share behavior. That creates a powerful opportunity for what we might call parasocial laundering. A message feels more trustworthy when it appears to come through a familiar face. Podcast hosts, musicians, actors, streamers, and influencers carry emotional credibility. Viewers may not trust a government account, a party account, or a faceless page, but they may trust the person they listen to every week while cooking, commuting, or falling asleep. If a clip travels through that familiar identity, the message can feel less like propaganda and more like a personal recommendation. Artificial intelligence intensifies this risk because it lowers the cost of imitation. AI can now generate voices, images, scripts, translations, captions, and increasingly realistic video. OpenAI reported in May 2024 that it disrupted five covert influence operations that tried to use its models for deceptive activity, although those campaigns did not appear to gain meaningful reach from OpenAI services. That caveat matters because the worst predictions about AI propaganda have not all materialized at scale. Still, the direction of travel is clear. AI is making content production cheaper, faster, and more adaptable. Meta said after the 2024 global election cycle that widespread AI deepfake disruption did not materialize in a significant way on its services. That finding should keep the conversation grounded. The danger is not always a single perfect deepfake that changes history overnight. The more realistic danger may be constant low-grade synthetic pressure, fake comments, fake personas, fake translations, fake local voices, fake screenshots, fake outrage, fake consensus, and fake uncertainty. Influence operations do not always need people to believe a falsehood completely. Sometimes they only need people to stop knowing what to trust. A convincing fake can deceive. A flood of questionable material can exhaust. A disputed clip can polarize. A manipulated recording can force institutions to spend precious time denying something. Even after a debunk, the residue remains. Some viewers remember the accusation more clearly than the correction. Short form video also changes the economics of narrative testing. In older media environments, propaganda required printing presses, broadcasters, formal networks, and slower feedback. In the clip environment, every post is a miniature focus group. Operators can see which emotional frame gains watch time, which caption triggers comments, which visual style attracts a demographic, and which influencer ecosystem is most receptive. That feedback loop allows narratives to evolve quickly. Again, the defensive point is not to provide a manner. The point is to understand why the environment rewards experimentation. The rise of state-linked content on short-form platforms shows that governments already understand the opportunity. Brooklyn's research in 2024 found that Russian state-affiliated accounts had begun leveraging TikTok's appeal as another avenue for disseminating overt Kremlin messaging, especially as the information ecosystem fractured across platforms with different identities. This is important because influence does not always hide. Sometimes the message is overt, labeled, and still effective because it is formatted for the platform. The most effective content often borrows the visual language of ordinary users. It may look like commentary from a bedroom, a stitched reaction, a fan edit, a street interview, or a podcast highlight. It may use subtitles, jump cuts, green screen reactions, split screen footage, and trending sounds. The format signals authenticity because it resembles everyday creator behavior. A polished government video may feel distant, while a rough clip with captions may feel immediate and real. This brings us to attention at the center of healthy use of technology. The same tools that can manipulate can also educate. Short clips can introduce people to complex ideas, historical speeches, safety information, medical guidance, language learning, financial literacy, and long form conversations they would never have found otherwise. A clip from a podcast can bring a listener into a serious interview. A movie clip can revive interest in older cinema. A song fragment can help an independent artist find an audience. A short video can help a young person feel less isolated. The US Surgeon General's advisory on social media and youth mental health acknowledges both benefits and risks. The advisory says social media use among young people is nearly universal, and it states that society cannot yet conclude social media is sufficiently safe for children and adolescents. It also recommends boundaries, tech free zones, responsible behavior, stronger privacy protections, independent research, and digital media literacy. The healthiest response is neither panic nor surrender. The healthiest response is design discipline, user agency, and social norms that protect attention. Common Sense Media and Hope Lab's 2024 report also points toward a nuanced picture. Among youth who use social media, YouTube, Instagram, TikTok, and Snapchat remain widely used, and online communication can feel safer and more supportive for some marginalized young people. That matters because technology is not experienced evenly. For some people, online spaces are a source of comparison, harassment, and compulsive use. For others, they are a source of belonging, discovery, and help. Healthy technology policy has to hold both truths at the same time. The question becomes how to preserve the good while reducing the manipulability. Hopefully that's the correct word. Anyway, that starts with understanding that attention is now infrastructure. A society's attention system shapes what people believe is urgent, normal, shameful, admirable, or impossible. Roads move bodies. Payment systems move money. Feeds move perception. When a society's Infrastructure is optimized mainly for engagement, influence operations receive a favorable climate. One healthy use of short form technology is intentional discovery. A viewer sees a clip, then follows it to the full podcast, the full song, the full film, the full report, or the original source. In that model, the clip is a doorway. The problem begins when the clip becomes the room. A person consumes fragments all day and mistakes familiarity for understanding. The more serious the subject, the more important the pathway from clip to context becomes. Another healthy use is creative participation. People remix songs, comment on films, respond to podcasts, and make meaning together. Culture has always been remixed culture. The difference now is scale, speed, and algorithmic amplification. A remix can be art, community, or criticism. It can also become manipulation when the audience cannot tell whether the remix is faithful, synthetic, paid, coordinated, or intentionally deceptive. A third healthy use is rapid public warning. Short video can spread safety information during disasters, expose abuse, document war crimes, and help people organize aid. During crisis, ordinary users can become witnesses. The danger is that the same crisis environment is also ideal for false claims. When people are afraid, angry, or grieving, they share faster. Influence operators understand that crisis compresses skepticism. Emerging technology is pushing the clip economy towards synthetic abundance. YouTube announced in 2025 that it was bringing a custom version of Google DeepMind's VO3 to YouTube Shorts, allowing users to generate video clips with sound from a phone. Google DeepMind describes VO3 as capable of generating native audio, including sound effects, ambient noise, and dialogue. Meta launched Vibes in 2025 as a feed for short-form AI-generated videos that users could create, remix, and cross-post to Instagram and Facebook stories and reels. This trend points toward a world where creation becomes conversational. A user will not need editing software, a camera crew, voice actors, or design skills to make persuasive media. They will describe a scene, upload a reference, choose a tone, and click publish. That could democratize creativity. It could help educators, independent filmmakers, small businesses, activists, and artists. It could also fill the feed with synthetic material optimized for whatever emotional response the system rewards. The next phase will likely be personalized persuasion. Today, most viral clips are broadcast to broad audiences even if algorithms segment distribution. Tomorrow's clips may be generated or edited for narrower psychological profiles. One viewer receives the security version of a message. Another receives the fairness version. Another receives the religious version. Another receives the anti-elite version. The core narrative remains consistent, but the emotional doorway changes. This is where psyop style risk becomes more serious. Traditional propaganda often struggled with audience diversity. A message that persuaded one group could alienate another. AI-assisted short form media may allow influence campaigns to produce many culturally specific variants at low cost. Translation, voice cloning, local slang, local grievances, and platform-specific aesthetics can all be adapted. The result is a persuasion environment that feels local, even when it is centrally coordinated. Society is already reacting through law, platform policy, parental behavior, and cultural fatigue. The European Union's Digital Services Act has pushed platforms to change systems and interfaces with the goal of creating a safer online experience for Europeans. Australia's under-16 social media restrictions came into effect in December 2025, and the eSafety Commissioner reported that age-restricted platforms removed access to 4.7 million under-16 accounts by mid-December. In the United States, TikTok's ownership and national security debates continued into 2026, with Reuters reporting a new U.S. joint venture structure intended to avoid a ban. These reactions show that societies are moving from content moderation debates into infrastructure debates. The question is no longer only whether one post should be removed. The question is how recommendation systems are designed, how age is verified, how synthetic media is labeled, how researchers access data, how political ads are archived, how creators disclose payment, and how users regain control over feeds that shape their attention. There will also be a cultural reaction against synthetic excess. As AI-generated clips become more common, some audiences will crave proof of humanness. Live events, long-form interviews, behind-the-scenes footage, local gatherings, and trusted hosts may become more valuable. The irony is that long form media may gain new importance as a trust anchor. A clip may introduce the idea, but the full conversation may become the place where credibility is judged. For podcast hosts, this creates both opportunity and responsibility. A show like mine, The Bait Picture, can use clips to reach people who would never search for a 40-minute episode. A strong clip can invite the audience into a deeper conversation. At the same time, the host has to assume that every clip can travel without the full episode. That means clip selection is editorial judgment. Titles, captions, thumbnails, and timestamps are not minor packaging decisions. They are part of the truth environment. For listeners, the key habit is context recovery. When a clip generates a strong emotional reaction, the next step should be simple. What is the original source? What happened before and after the clip? Who posted it first? Is the account authentic? Is the audio original? Is the caption adding a claim that the video itself does not prove? Has the person in the clip responded? Are reputable outlets confirming the event? These questions do not require paranoia, they require friction. For platforms, the key design challenge is slowing down the most dangerous forms of frictionless amplification without destroying creativity. Labels can help, but labels alone will not solve the problem. Watermarks can help, but watermarks can be removed or ignored. Provenance standards can help, but only if they become easy for normal users to understand. Rate limits, virality circuit breakers, add transparency, researcher access, and stronger penalties for coordinated deception may become more important than individual takedowns. For governments, the challenge is legitimacy. A government that responds to manipulation with vague censorship will deepen distrust. A government that ignores manipulation will leave the public vulnerable. The better path is transparency, narrow rules, independent oversight, media literacy, and public evidence. Democracies should be especially careful here because the cure can damage the patient. A society cannot defend trust by teaching people that institutions can quietly decide what they are allowed to see. For creators, the challenge is incentives. The clip economy rewards outrage, certainty, humiliation, and speed. A creator who refuses those incentives may grow more slowly but may build a more durable audience. The practical question is whether creators can make context compelling. Can they produce clips that attract attention without misleading? Can they turn curiosity into full listening rather than fragment addiction? Can they make correction visible when a clip travels badly? There is also a business dimension. Short clips are now part of the funnel for podcast, music, film, television, news, shopping, politics, and ideology. That means the same tactics used for marketing can be adapted for manipulation. A campaign can test hooks, segment audiences, recruit micro influencers, and optimize messaging. The moral difference is purpose, transparency, and truthfulness. The technical machinery is often similar. So the future may divide media into three broad categories. The first category will be human-made media with strong provenance. The second will be openly synthetic media used for creativity, entertainment, education, and experimentation. The third will be ambiguous media, where origin is unclear and trust depends on social signals. The third category is where influence operations will try to live because ambiguity is useful. It allows believers to believe, skeptics to argue, and manipulators to move on before accountability arrives. A healthier technology culture would make ambiguity less profitable. That does not require burning creativity. It requires better defaults. Synthetic voices should be disclosed when used in public persuasion. Political content should carry stronger provenance signals. Platforms should give users meaningful control over recommendation systems. Schools should teach source tracing as a basic civic skill. Podcast hosts and creators should publish full context and correct misleading clips quickly. The audience also has power. People can reward creators who provide context. They can resist sharing clips that confirm their anger too perfectly. They can treat virality as a reason to inspect, not a reason to obey. They can remember that emotional force is not the same as evidentiary strength. The feed trains people to react. A healthy media citizen trains themselves to pause. The biggest misunderstanding about short-form influence is that manipulation only works on foolish people. That belief is comforting, but it is false. Smart people are vulnerable to messages that flatter their identity, confirm their suspicions, or arrive through trusted cultural channels. Education helps, but identity can overpower education. The most dangerous clip is often the one that makes the viewer feel clear, righteous, and socially reinforced before they've checked anything. Another misunderstanding is that truth always wins if given enough time. Truth often does better with time, but attention has windows. Elections, protests, riots, bank runs, boycotts, military escalations, public health scares, and reputational attacks can all move faster than institutional verification. A false clip released at the right moment can shape behavior before the correction arrives. The correction may be accurate and still be late. This is why the Slovakia episode remains important even though scholars caution against simplistic conclusions. The deep fake did not operate in a vacuum. It entered a low trust environment with pre-existing polarization and geopolitical tension. That is how influence usually works. It exploits cracks already present in the social structure. Technology accelerates the spread, but the vulnerability begins with distrust. The hopeful part is that resilience also begins before the crisis. A society with trusted local journalism, transparent institutions, media literacy, healthy online norms, and accountable platforms is harder to manipulate. A podcast audience trained to expect nuance is harder to radicalize through fragments. A creator ecosystem that values correction is harder to weaponize. A young audience taught to trace originals is harder to hurt. The clip economy is not going away. YouTube Short Alone was reported by YouTube CEO to average more than 200 billion daily views in 2025, nearly tripling from 70 billion daily views in March 2024. Pew found that 90% of US teens use YouTube, while majorities use TikTok, Instagram, and Snapchat. Short form video is now a default layer of youth culture, entertainment discovery, news exposure, and public argument. The goal, therefore, is not nostalgia for a slower media world. The goal is a more disciplined, fast media world. We need short clips that point toward context, platforms that reduce deceptive amplification, creators who understand editorial responsibility, and audiences who treat emotional certainty as a signal to investigate. We also need institutions that communicate quickly without becoming reckless, because silence leaves a vacuum that bad actors can feel. The rise of podcast clips, song clips, and movie clips tells us something profound about modern persuasion. People are not only persuaded by claims, they are persuaded by voices, moods, scenes, rhythms, jokes, faces, and feelings of belonging. A political argument may fail as a speech and succeed as a sound. A conspiracy may fail as an essay and succeed as a montage. A foreign narrative may fail as a press release and succeed as a creator style video. That is the bead picture. The clip is now a cultural molecule. It is small enough to move quickly, emotional enough to stick, and flexible enough to be recombined. In healthy hands, it can invite learning, creativity, discovery, and connection. In manipulative hands, it can launder propaganda through entertainment, manufacture social proof, and turn uncertainty into a weapon. The future of media literacy will not be only about spotting fake images. It will be about understanding how fragments shape perception. It will be about asking why this clip reached me, what feeling it wants from me, what context it removed, and who benefits if I share it right now. The next era of Psy Up Resilience will require more than fact-checking. It will require attention discipline. The most important defense may be a simple habit. When a clip makes you instantly furious, instantly smug, instantly afraid, or instantly certain. Pause long enough to recover the larger frame. The feed wants reaction at the speed of the thumb. A healthy mind can still choose the speed of judgment.