By Wayne Dominowski, SBA Guest Commentary
(SB) – Artificial Intelligence (AI) is upon us, and I am not sure it is altogether something we can applaud. I’ve already witnessed fake YouTube stories. One was of a ‘Catholic priest’ singing a sad song (with a message) in a supposed talent competition. It’s one of many.
A video of an aging rocker named Michael Bennett performing a heart-tugging song on America’s Got Talent elicited reactions around the globe. But the heavily tattooed musician with long white hair and thick silver rings on his fingers isn’t real. He’s AI-generated.
In a clip seen by millions, a cute little house cat saves a baby from a charging bear. It’s the stuff of viral video gold: drama, innocence, animals, peril - all engineered for the millions of clicks it racked up on Facebook and Instagram.
Artificial Intelligence is one thing, daily news, both printed and on-the-air, have infiltrated the national landscape. From infiltration to common use, the fake stories, fake news, fake reports, and fake videos are part and parcel today.
Most people are aware of this. Those who are have become skeptical of what they hear and see.
Psychology Today reports that “Fake news is more widespread today than ever before.
Political activists who want to unfairly influence voters create false news stories. Unscrupulous media companies publish falsehoods as clickbait. Political regimes spread false propaganda to support their policies. And the blurring line between entertainment and news leads to writing that doesn’t comply with journalistic standards.
The proliferation of fake news has led to a new body of research on the topic: Since the 2016 U.S. presidential election, more than 4,000 scientific articles have been published on the topic of false news. A systematic review published last month asks the question, how good are people at judging the accuracy of the news?”
How can you tell if a video is real or AI? Jennifer Jolly, writing for USA TODAY, offered the following precautions:
1. Check the source. If it’s real, multiple credible outlets or witnesses will confirm it. If every link leads back to the same viral TikTok post, then it’s probably not real. No, a random “expert” on YouTube is not as credible source.
2. See a “spongy” spot? Many fakes are edited to erase the original watermark. Look closely near the corners or edges of the frame for a blurry patch, smeared light, or a soft square that looks slightly out of focus - that’s where the watermark used to be.
3. Watch the camera, not the subject.AI is great at faces, terrible at filming them. Real video wobbles slightly with the person holding the phone. AI clips move too smoothly, as if they’re gliding on rails. If it feels more like a drone shot than a phone shot, pause before you believe it.
4. Check the physics. Look for light and shadow behaving naturally. Does a sleeve flutter when the wind blows? Do reflections reflect? AI often gets these almost right - but not quite.
5. Listen closely. Footsteps or voices might drift out of sync, or the background hum feels too clean - as if someone hit “mute” on reality.
6. Study the account. If a video comes from a mystery handle with no bio, no history and no location, assume it’s fiction until proven otherwise.
7. Note the format. Fake “bodycam” and “security” footage videos are exploding because that grainy chaos hides AI’s flaws. If it looks as if it came from a cop’s chest cam or a Ring doorbell, question it.
8. Slow it down. Going frame by frame often highlights telltale AI hiccups such as a melting logo, a vanishing hand, jumbled text on signs or text that flickers and changes.
Assume every video you see on social media is fake from now on, until credible media outlets cover it.
You don’t have to spot every fake - just slow down enough to question what you’re seeing. In 2025, skepticism isn’t cynicism, it’s self-defense.
What happens next?
OpenAI says it’s building stronger labeling systems and partnerships with platforms to trace Sora clips back to their source. It’s also giving people control over how their likeness can be used - though copyright enforcement still depends on rights holders filing complaints.
Google just rolled out its newest AI-video generation model, Veo 3.1, Meta is racing to show off its own social AI video platform, and others will soon follow.
The race isn’t just about who makes the most realistic clips; it’s really about who controls the attention they command.
For now, skepticism is the new literacy. Look twice, think all the time, and remember: Your intuition is the watermark they can’t erase.
ED. NOTE. Jennifer Jolly is an Emmy Award-winning consumer tech columnist and on-air contributor for “The Today Show.”
