AI or Real? How you can spot real content versus AI-manipulated fakes

This investigative report contains AI-generated images and videos created to show you the best ways to identify real versus AI.

This is the first part of InvestigateTV’s mAnIpulated series, examining the ways in which AI is impacting our everyday lives.

Visit the series homepage to follow each national release as well as reports from your local Gray Media stations.

(InvestigateTV) — Content generated and manipulated with the assistance of Artificial Intelligence (AI) is becoming indistinguishable from reality.

Could you spot an AI-manipulated fake?

Take a look at these two images of InvestigateTV’s Kristin Crowley.

Do you see the difference?

While Kristin was actually standing in Los Angeles when this footage was captured, the original image on the right was manipulated using an AI tool to create the new image on the left. This AI image manipulation allowed Kristin to look as if she were reporting from directly in front of the well-known Hollywood sign.

Famous places and people are becoming easier to fake as AI technology advances each year.

Misinformation can spread with something seemingly small, perhaps a fake story or photo of a major news event. Maybe even a fake image of a famous person is used in the story as well.

This online deception can seem harmless at first, like with AI-generated videos of bunnies jumping on trampolines, but what if the posts don’t stop at that point? The result can be scams, community discord or chaos spread via false information.

Some content creators are using their powers for good. Instead of deception tactics, they are turning the tables on the misinformers to educate people on how easy it can be to get duped.

Our investigators traveled to California to meet with two creators who are equipping people with the skills to spot fakes by creating their own AI-powered content on popular social media apps and video streaming sites.

AI or Real?

When our investigators met Madeline Salazar, she was on a grocery run in San Diego.

Her shopping list was short, and the items were not being purchased for dinner.

Salazar creates videos showing off the power of AI to her social media followers.

In a popular example, she is holding what looks like two purses. Then, in the reveal, the purse on the left is actually a potato, and the purse on the right is the real thing.

“The potato was just a very easy thing to hold and to motion track, and people thought it was really funny,” Salazar said with a laugh. “So, I’ve leaned into it, and now I pretty much have a potato with me wherever I go.”

The potato idea seems to be working for her videos.

When asked how successful the AI or Real series has been, she said the views reach into the millions, with some reaching as many as 15 million views each.

Salazar believes the high view counts prove her viewers have a high appetite to learn more about AI.

Salazar walked our team through programs she uses to create her videos with the help of AI, including Adobe Firefly and Photoshop.

“I take a still from my video and bring that into Photoshop, where I will generate whatever AI image I would like to show,” she described. “I can lay that graphic on top of my footage, and you won’t be able to tell the difference.”

While she thinks an average, inexperienced person would find creating videos like hers challenging, Salazar knows the intelligence behind the technology is only going to advance.

“The average person probably can’t create an incredible, indecipherable AI-generated video and post it, but it is progressing every day.”

Salazar hears concerns about her content almost daily.

“A lot of people are very upset about the fact that I’m teaching other people how to do this,” she said. “But in my opinion, I am showing you how to be aware of it.”

She believes many push back on her videos because those creators who are taking advantage can get large amounts of views in short periods using AI-manipulated content.

“They know they can go viral by making a generative AI video of a natural disaster in someone’s city, and that’s terrifying.”

“It’s Not Coming. It’s Here.”

Travis Bible has a background in film as a director and an editor. He created a public service announcement video for his parents to show them exactly what AI could do.

That video generated over 2,000,000 likes on Instagram.

“I’m trying to create AI awareness,” Bible told InvestigateTV. “I don’t want people to be caught off guard, kind of like the way I was, with how far this technology has come. It’s not coming, it’s here.”

In some ways, Bible says AI has a long way to go.

“I did a video comparing all-time great human-actual performances in AI, tried to replicate them, and it was a train wreck. I did it to remind people how important these actors and humans are to creative levels.”

Bible walked our team through some common AI flubs people can look out for to help spot otherwise convincing AI.

“We could go through some things, how it looks too professional, like there’s some blurring around the edges. If you look at the background, weird things are happening, but a year from now, that might not be true anymore. That’s how fast it’s advancing.”

Better Tech, Bigger Risk

The Department of Homeland Security released a report warning that deepfakes pose threats to national security, personal finances, and more.

The report’s findings included the following:

• The effectiveness of deepfakes in spreading misinformation doesn’t rely only on advanced technology, due to people’s tendency to believe what they see.
• Expert opinions on the urgency of deepfake threats vary widely from “urgent” to “be prepared.”
• The risk of deepfake attacks is expected to rise as the cost and resources needed to produce them decrease
• Combating deepfakes requires a multi-pronged approach involving innovation, education and regulation.

Read the full Department of Homeland Security report:

Resemble AI, a voice technology company, issued a report this year detailing that AI-powered deepfakes caused more than $200 million in financial losses in just the first quarter of 2025.

The report analyzed over 160 documented deepfake incidents occurring between January and April 2025.

The report’s key findings included:

• Women, children and educational institutions face growing deepfake threats
• Criminal exploitation has evolved beyond scams and now includes targeted harassment and blackmail schemes
• Incidents are impacting both developed and developing nations across the world

Read the full Resemble AI report:

Common Sense and Critical Thinking

Travis Bible says critical thinking and common sense can be much more trustworthy than our eyes and ears.

“Do your best to use common sense on if this would actually be happening. Because the line between what’s AI and what’s real and how you can tell, it’s getting very blurred.”

In some cases, even deepfakes labeled as AI have still tricked people. It recently happened to television host Chris Cuomo.

He reposted this deepfake video of Congresswoman Alexandria Ocasio-Cortez. The video was already labeled as being an AI creation.

Ocasio-Cortez responded on X, reiterating that the video was a deepfake.

Cuomo acknowledged his error and deleted the post.

Bible believes these AI fakes, from those that seem harmless to those disseminating misinformation, can create lasting repercussions.

“You’re starting to erode what’s true and what’s fake, and I think just little things like that can kind of, those innocuous things can kind of lead to more and more of this stuff and people not knowing what to believe.”

That’s why Travis Bible and Madeline Salazar continue creating their examples, showing the “real side” of artificial reality.

Put Your Skills to the Test

Can you spot the AI-manipulated images in our new interactive digital game below after the lessons Madeline Salazar and Travis Bible taught in this story? You can play this game and more at our mAnIpulated series homepage, as well.

Check Out the mAnIpulated: A mIsinformAtIon nAtIon Series from InvestigateTV

This is the first story of InvestigateTV’s mAnIpulated: A mIsinformAtIon nAtIon series. Check the series homepage for updates with the latest reporting on AI-related topics from our local Gray Media stations throughout August and September.

Watch & Read More from InvestigateTV

Sign up for the InvestigateTV Newsletter, subscribe to our YouTube channel and much more with the toolbox below featuring our go-to links.

Copyright 2025 Gray Media Group, Inc. All rights reserved.