If you haven't heard, GPT released something called Sora, which is a video generation tool similar to GPT, but instead of it being text or images, it's now video. If I'm sure you've seen one or two videos online that friends have shown you, saying, "This isn't real." Now expect that number of videos to triple in 2025.
There used to be ways to figure out whether images were real or not. You could look at people's eyes or fingers and realize, "Oh, okay, they have seven fingers, so they're not real," or notice they looked pixelated. But all of that has been fixed. Tools like MidJourney, for example, create images so realistic that you cannot tell whether they're real or not. There was a famous image of Donald Trump that was fake, although no one actually knew it was fake. That image forced MidJourney to cancel free accounts to limit such misuse.
The Challenge of AI-Generated Videos: Google's Veo 2 and the Disruption of Reality
Now, videos are a totally different challenge. At the moment, most AI-generated videos have problems with physics, similar to how AI had problems with fingers in images. But it has been said that Google's latest video generation tool, Veo 2, doesn't have these physics problems or has solved them. This gives Google an edge in the AI video race and, most importantly, disrupts our perception of reality.
Face it: you're going to be looking at your feeds and will not be able to distinguish between AI-generated and real videos. You know how, when you see a picture of someone, you sometimes wonder whether it's Photoshopped? A little retouching, enhancing the lighting, and other tweaks can make it hard to tell. With videos, there's not even a way to find out because the output will be flawless.
The New Normal: Accepting the Ambiguity of AI-Generated Content
This is going to be our reality in the upcoming years. You're going to watch a video, wonder whether it's AI or not, and give up trying to figure it out. Then you'll move on. This shift could open up a huge opportunity for the offline world. Believe it or not, face-to-face meetings might become more valued as a way to verify authenticity. Validation and trust might increasingly require physical presence.
The Implications of AI-Generated Videos: Conflicting Information and the Need for Verification
Imagine you're looking forward to watching a football game. You check YouTube for highlights and see a video showing one team winning. You tell your friends, only to find out they watched another AI-generated video showing a different score. The real score? You'd have to check verified news sources to find out—but even those might be manipulated to optimize for search engines already overwhelmed by AI content.
Tough times are ahead, so we'd better open our eyes. The blending of reality and AI is about to change the way we perceive information and interact with the digital world.