You are scrolling through your feed when a screenshot appears showing a public figure saying something surprising or controversial. Within minutes, it is everywhere. Some are angry, others defend it, memes parody it, and arguments spread across platforms.
Later, you discover the person never quite said that exact quote. The words came from a longer interview, the clip was shortened or an incorrect caption was added. But the screenshot has travelled faster and further than the original video ever did. What people reacted to was a version of the message created through circulation, rather than the message as originally delivered. Sound familiar? This pattern can be seen across nearly every viral moment, from political speeches to celebrity interviews.
Research in media and communication studies has long shown that meaning rarely remains fixed once a message enters circulation.
My work examines how these small shifts accumulate as messages move through digital environments. I describe this process as “message drift”, where content becomes separated from its original context as it is clipped, reposted and reframed across digital platforms.
Message drift rarely occurs in a single dramatic moment. Rather, it emerges through a series of small transformations. Evidence suggests this is driven both by limited user attention, and by platforms that prioritise content which is quick to consume and easy to share. Studies of digital attention show that people engage with information in short, fragmented bursts. Social media platforms tend to amplify content that is easily processed and widely circulated.
My research suggests that audiences interpret statements through contextual cues such as captions, commentary and surrounding visuals. The same words can appear sarcastic, alarming or definitive depending on how they are presented.
During the defamation trial between actors Johnny Depp and Amber Heard, short clips and reaction edits spread rapidly across TikTok and other platforms, with viewers often forming strong opinions from fragments of much longer courtroom proceedings. Different clips, captions and reposts pushed audiences toward opposing interpretations of the same events. Fan culture on social media further intensified online reactions around the case.
The circulation of AI-generated images during the 2024 US election cycle also demonstrated how quickly visual content can detach from its original context. Edited campaign images, parody posts and reposted screenshots spread rapidly across platforms and were repeatedly mistaken for authentic political material. Fabricated visuals were shared and debated at speed before people could verify where they came from or whether they were real.
What is AI slop? Why you are seeing more fake photos and videos in your social media feeds
Short clips of political speeches are frequently shared with captions that emphasise a particular interpretation, even when the full speech suggests a different tone or meaning. For instance, in 2021, images of a crowded House of Commons circulated online with captions claiming MPs were debating their own pay. The image was used to suggest MPs were paying more attention to their own salaries than to other political issues. In reality, at least one of the images was taken during a different debate and mislabelled.
Andrii Iemelianenko/Shutterstock
In older media environments, such as broadcast television, newspapers or official press releases, authority depended on who said something and where it appeared. Online, authority increasingly depends on repetition. The version of a message encountered first, or most frequently over time, often becomes the version people treat as the most accurate or authoritative account.
In my own ongoing research, I have asked participants to respond to edited clips, screenshots and images to test how they interpret messages in different formats. When these differing versions conflict, participants have defended their initial interpretations and challenged others, causing disagreement to escalate into argument.
Why this matters
Conflict is not always about competing interpretations of a single statement, where people disagree about what the same message means. Instead, it often arises because people encountered different versions of that statement in the first place.
Digital platforms themselves further intensify this. Studies of online behaviour show that emotionally engaging content is more likely to attract attention and be shared. These versions are more likely to dominate what audiences encounter. As a result, simplified or emotionally charged versions of messages often travel further than the original. It is quicker to process, easier to repeat and more likely to trigger emotional reactions, all of which are linked to higher engagement.
Over time, this can contribute to wider problems in online communication, including polarisation and misinformation. New digital tools, such as generative AI, make it easier to produce realistic but altered images, video and audio at speed and scale, increasing the likelihood that misleading or context-shifted content can circulate widely.
Next time you come across a clip that you have the urge to share or comment on, ask yourself a few key questions. Where did this message first appear? What context might be missing? And how many different versions of it might now be circulating?
Understanding message drift does not resolve these problems on its own, but it can help people recognise what is happening when online debates escalate.
