AI is tricking the best of us. Here are 7 dead giveaways that it's fake.
Recently, there have been a string of videos depicting people, mostly women, becoming overcome with fits of rage in grocery stores and restaurants over Supplemental Nutritional Assistance Program (SNAP) benefits being cut. The videos have been shared thousands of times as a means to shame the people in them for their bad behavior, while simultaneously upholding the stereotype around who receives SNAP benefits.
Comment after comment tells the women in the videos to get jobs and stop relying on government assistance to feed their children. The only problem is, the women aren't real. The videos were generated with Sora and other Artificial Intelligence (AI) apps that allow users to create realistic AI videos with a quick description and a couple of clicks. While some people were quickly able to identify the problematic videos as AI, a lot of people couldn't.
As AI technology continues to improve, it becomes even more difficult to identify when a video or picture is real, but there are some easy tells people can look for.
1. Their skin has no texture
No one's skin is perfect, except if you're a computer-generated depiction of a human. Everyone else has texture, blemishes, fine lines, or bags under their eyes. The reason for that is being human is rough sometimes, and it shows on your face. Some people might have acne scarring, freckles that aren't perfectly placed, or a night of terrible sleep that shows in their eyes. People also tend to move their faces throughout their lifetime, which causes laugh lines and crow's feet from squinting. We have pores, too.
If the person in the picture or video doesn't appear to have skin with depth, texture, or other imperfections, chances are, they're not a person at all. According to the BBC, "AI images often lack details that can be found in real pictures, leading to these photos having an 'airbrushed' look."
Portrait of a woman in a serene, natural setting.Photo credit: Canva (AI-generated)
2. The words aren't words at all
This one can be easily overlooked because of brand recognition. AI will use a familiar logo, which will trick your brain into creating the words around the logo because you know what it's supposed to say. But take an extra few seconds to read the words in the picture or video because, if it's AI-generated, the words will be gobbledegook. Just a hodgepodge of random letters and symbols that mean nothing.
Interview in the grocery store aisle.Photo credit: Canva (AI-generated)
3. Look for anomalies that don't make sense
Of course, humans are born with all sorts of anomalies, but they are generally differences that we are accustomed to seeing. These are things like a small extra digit on the side of someone's hand, or one leg that might be a little shorter than the other. But when AI is creating full-body pictures, something big but easily overlooked is almost always off, like a human leg growing out of a dog's neck in the family picture below. Or a hand that looks like it belongs to a woman attached to a man's arm turned completely the wrong way, while the other hand looks normal.
From a distance or at first glance, someone could easily overlook the discrepancies in photos or videos, so it takes looking a little longer to clock them.
Group hug with furry friends in the park! 🐶📸Photo credit: Canva (AI-generated)
4. Things blend or are blurry
One thing that will give AI away is blurry mouths while the person is speaking. Their lips and sometimes their tongue, will become blurry, pixelated, or even slightly disappear.
Sarah Kimmel, from Family Tech Zone tells KSL News Utah, "AI has a really hard time, especially with hands and transitions, so you know, my shirt ends here but with an AI image, it's going to blend into the background more or my belt will blend into my shirt, or my hair is going to blend into the background as well so you want to pay attention to those sharp edges and also, whether or not the picture is even feasible. So there's these beds on the side of a cliff, and you're like, 'How does somebody even get to that bed?' That image is not possible, it's AI-generated."
5. Syncing is slightly off
This isn't just something you'll notice with mouths as the AI-generated people are speaking, but other things will be a little off, too. A video of someone running a bath may show the bath water turning on just before the person reaches the tub or after they've already walked away. You might hear a door close right before the person closes the door.
6. It makes you unexplainably uncomfortable
Humans are hard-wired to connect to other humans, even if it's through a screen. Watching a computer pretend to be human can cause little alarm bells to go off in your head or belly that tell you that something isn't quite right. Everyone can't put their finger on what it is, but our body often gives clues that the video you're watching is "not one of us." Hopefully, we don't have to walk around like "Among Us" characters anytime soon.
Ruben Circelli from PC Mag writes, "Listen closely to a video’s audio, and pay special attention to timbre, which is the character of a sound independent of its pitch or volume. AI-generated voices, in particular, often have a vaguely robotic timbre. And even if you can’t describe exactly why they sound off, it’s usually easy to hear when something isn’t quite right."
7. The original source is sketchy
When all else fails, when something feels off, you can take a screenshot of the video or picture and run it through Google's reverse image search. This will help bring up the source of the content. If it's AI, it will likely bring you to a dead end or take you to a social media page that only contains other videos or images that also seem off.
Going to the source before purchasing something from an ad online can also save you frustration. People use AI to generate products as well, so going to your browser to type in the company name, like Barbie or Levi's, to search for the product is one way to beat getting duped by AI.
At the heart of identifying AI content is critical thinking. Taking a critical look at the content in front of you will save you from sharing something inaccurate. While the AI videos are supposed to be used for fun, some bad actors are using them to spread misinformation or create anger and fear. In the age of AI, it's important to slow down, do a quick Internet search of the video to see if a reputable source is discussing it, or look for a few of these easy giveaways.