Most people think they can recognize fake news. But today, the problem isn’t just simply identifying what’s fake, but realizing how much of what we read was never real to begin with. 

Take this: Google once claimed that a small town in South Wales, Cwmbran, had the highest concentration of roundabouts in the world—a weirdly specific statistic but highly believable without much question. The truth, however, is far less impressive. The claim originated from an April Fools’ joke written by Ben Black, a 48-year-old local journalist who runs a community news site. What was clearly labeled as an “April Fools’ Day joke” was later picked up, repeated, and ultimately misinterpreted by AI as a factual, true statement. A joke, stripped of its context, became “truth” simply because it was processed by none other than artificial intelligence. 

Mr. Black, who had built a tradition of publishing playful stories each year—from fictional landmarks to absurd local events—had never intended for his writing to be taken seriously. Yet, when he searched for his article later, he found it already taken on a life of its own. Google’s new AI-generated response displayed the claim as true, being the first thing that pops up when you Google search “roads in Wales.”

And while this example might seem harmless, it shows how easily fake information can spread — and what happens when it doesn’t stay as harmless and happy. 

In fact, consequences become far more serious when misinformation intersects with real-world events—and it’s happening more than you think it is. During the recent conflict with the US and Israel launching strikes on Iran, AI-generated videos depicting missile strikes circulated across social media, despite being completely false and created by artificial intelligence. Some clips showed cities under attack with fake missile alarms, buildings engulfed in flames, and civilians fleeing in panic. Many AI videos even showed false footage of Dubai’s Burj Khalifa skyscraper in flames, with a crowd of people running towards the building. Of course, these videos erupted all over the internet. People viewed them, and people shared them, eventually reaching millions. In some cases, AI tools even incorrectly verified these videos as real, further blurring the boundaries between fact and fake.

Now this is where the issue shifts from concerning to dangerous, because misinformation does not stay in isolation—it shapes perception, influences behaviors, and ultimately alters how people understand reality. When something appears convincing enough, people no longer pause to question it but react instantly. Panic spreads, false narratives form, and once something has been repeated enough times, it’s never leaving anyone. Specifically in wartime, these videos can cause emergency lockdowns from citizens who flee for the fear of their families’ safety, allowing unresearched news outlets to spread this false breaking news and putting panic and fear in people just because of this AI misinformation. And while this is just one example, there are tens and thousands of more examples out there. 

And the longer this continues, the more that uncertainty becomes a part of how we think. Because eventually, the problem is no longer just misinformation existing, but about people losing their confidence in their ability to recognize it. If we don’t stop this now—in the future—when everything looks believable, we will never be able to tell what is true and what is not. 

Here’s the thing: AI itself isn’t the problem. It’s how easily we consume it. Because the issue isn’t that false information exists—but it’s that it no longer looks false. And once that line disappears, it becomes much easier to accept things without questioning them.

So next time you come across something online that feels believable, don’t automatically pass it on. Pause, and ask yourself where it came from — and whether it was ever real to begin with. Because in a world where a joke about roundabouts can turn into a fact, and videos can create events that never happened, recognizing the truth is no longer automatic. 

It’s something you have to choose to do.