The world of AI-generated videos is a tricky one, and it's time to shine a light on this growing issue.
Imagine you're relaxing after a hearty holiday feast, and your uncle starts scrolling through a feed of vertical videos. He asks, "Did you see the one with the cat grabbing a snake from a guy's bed?" Your head starts to pound as you wonder, is it real or fake?
"We're drowning in a sea of slop," warns Mike Caulfield, co-author of 'Verified: How to Think Straight, Get Duped Less, and Make Better Decisions about What to Believe Online.' It's overwhelming, and our minds can only take so much before they give up.
But here's the good news: you don't have to surrender to despair just yet. There are simple strategies to help you navigate the authenticity of online content.
Don't Assume Everything is Fake:
With so much 'slop' in our feeds, it's tempting to think everything is fake. However, Kolina Koltai, a senior investigator at Bellingcat, warns that this bias is just as dangerous as believing everything is real.
Bystander videos are crucial evidence for uncovering misdeeds, and when people stop trusting these videos, it becomes easier for bad actors to claim real events are fake, a phenomenon researchers call the 'liar's dividend.'
Koltai emphasizes, "The bigger risk is not believing real videos."
Pay Attention to Video Features:
AI-generated videos are impressive and rapidly improving, says Hany Farid, a professor at UC Berkeley who studies manipulated media. Even experts can be fooled.
The key tip? Video length. Most AI videos are limited in duration because creating them is computationally expensive. Many are just 8-10 seconds long. While longer videos can be pieced together, these short clips are a good indication to pause and reflect.
AI videos also tend to perfectly frame their subjects, with main characters prominently featured and actions starting and stopping cleanly. This is evident in the quiz video of the NYC police officer shouting at ICE agents - it has a professional look to it.
Check the Context:
The place where a video is shared can be even more telling than its features, says Caulfield.
Checking the source and reading the comments can provide valuable clues. For example, the second ICE raid video in the quiz came from a Reddit community specific to Chicago's Logan Square neighborhood.
Similarly, consider the poster's feed. If they have diverse content beyond just immigration raids, it adds credibility to their claim of witnessing the event.
A reverse image search on Google can also help trace the video's origin, find other videos from the same event, or locate news reports confirming or refuting it.
Don't Feel Pressured to Share:
In an era where algorithms prioritize speed over accuracy, sharing isn't always caring.
Much AI content is engagement bait, says Koltai. Its creators often have a financial incentive to get you to like, comment, and share, as it can lead to more money for them.
When in doubt, Caulfield advises waiting. "You don't have to be the first to share; you can be the one who waits." Often, within hours, videos of events are confirmed by other sources.
The Impact of Sharing:
Many might think it's harmless to share an AI video of bunnies on a trampoline or a cat grabbing a snake, but experts disagree. When people are duped by AI videos, it undermines faith in the videos that truly matter.
"People wonder if it's really a big deal," says Koltai. "But we should all be worried. If we can't tell what's real and unreal online, it's incredibly dangerous."
Hany Farid agrees, "Every like, click, share, and engagement makes you part of the problem."
So, the next time you encounter a questionable video, remember these tips. Your critical thinking skills are the best defense against AI-generated slop.
