Sat, Jul 13, 2024
AI-generated photos flood Facebook to drive engagement

AI-generated photos flood Facebook to drive engagement

  • PublishedJune 13, 2024

Striking AI-generated images of service members, quadruplets or children in rubble are often shared on Facebook, and they could be used as part of a scam tactic.

If you’ve spent any time on Facebook in recent months, you have probably seen photos shared with this caption: “Why don’t pictures like this ever trend?”

VERIFY found multiple examples of these types of viral posts, including photos of children in rubble, quadruplets celebrating a birthday or military service members holding American flags while on a semitruck. All of the photos amassed thousands – or even tens of thousands – of likes and shares on Facebook. 

Kathy sent us this example of one of these types of posts to VERIFY and asked us if the photo is real. 

VERIFY looked into whether the photos shared in these Facebook posts are legitimate. 


Are these eye-catching Facebook photos with captions like “why don’t pictures like this ever trend” real?



This is false.

No, these photos are not real. They are AI-generated images. 


VERIFY analyzed three different images shared on Facebook with the caption “why don’t pictures like this ever trend?,” including Kathy’s example, and found they were all created using artificial intelligence. Generative AI is a type of artificial intelligence technology that can produce new content, including images, videos, audio and text based on a prompt.

While some Facebook pages that share this AI-generated content may be harmless, others could be trying to scam people. 

Some of the posts may be an example of a scam technique called “like-farming,” where scammers create posts that are designed to get engagement such as likes and shares. These posts could include anything from cute animals to political messages, the Better Business Bureau (BBB) explained in 2020.

The posts or Facebook pages themselves may end up promoting fake products or directing people to outside websites that could be malicious. 

To confirm the three images shared on Facebook shared captions were created with AI, VERIFY first conducted a reserve image search and found no other reputable sources sharing the same photos online. Hive, an AI detection tool, also said it’s extremely likely all four of the images are AI-generated. 

Then, we looked for red flags that point to the images being created with AI. Here’s what we found for each one. 

When looking at the image that appears to show children in rubble, we found telltale signs of AI.

Generative AI is notoriously bad at getting hands and fingers right. The hands in the AI-generated image of the children in rubble are misshapen, with some of them missing fingers and others having too many. 

For example, the child on the left-hand side of the image only has four fingers and the hand on his shoulder has an extra thumb.

There are other inconsistencies in the photo, too. The arm that’s around the boy on the left-hand side of the image has a sleeve matching the color of his own shirt, instead of a color matching one of the other children’s shirts. 

The photo caption also has nothing to do with the image itself since it refers to a “beautiful cabin crew” and actress Scarlett Johansson.

Another AI-generated photo that’s made the rounds on Facebook claims to show quadruplets who turned 101 years old. 

One of the first steps to figuring out whether an image is AI-generated is checking the context. You should ask yourself if what you’re looking at is believable.

The people in this photo do not look 101 years old, as the caption claims. And, if quadruplets really did turn 101, news outlets would be reporting on it. But VERIFY could not find any legitimate news stories about 101-year-old quadruplets. 

You should also check the photo credit. On Facebook for mobile, this photo has Meta’s “Made with AI” label

At first glance, it might be difficult to spot some of the other red flags that point to this image being AI-generated. But, if you zoom in, you can find them. 

A remnant of a person’s hand is in the photo where it shouldn’t be and the person on the left side of the photo has a distorted hand, with texture that looks like a repeating pattern. 

AI also struggles to generate realistic teeth and ears. The teeth in this photo appear to be too small and one of the women in the photo has two different earrings. 

One of the women is also wearing a watch that isn’t consistent with what one would actually look like. 

Military service members with American flags on a semitruck

The AI-generated image of military service members holding American flags on a semitruck that Kathy sent to VERIFY is another reminder to check the context.

People wouldn’t be able to stand on an open semi-truck trailer the way they are in this image. The semi-truck in this AI-generated image also has too many tires. 

But there are other AI red flags. For example, the semi-truck’s license plate and the service members’ bodies are distorted.

The license plate doesn’t show any actual numbers or letters. In some areas, the service members’ bodies blend together.

AI also struggled with texture here, since the road looks far too smooth.

How “like-farming” scams work and tips to avoid them

The like-farming scam can start with an initially harmless, albeit fictional, post. But when the scammer gets enough likes and shares, they will edit the post and add something like a malicious link that downloads malware to your computer, according to the BBB. In other cases, scammers may strip the page of its original content and “use it to promote spammy products,” the BBB says. 

Researchers at Georgetown and Stanford Universities studied more than 100 Facebook pages that regularly post AI content, and found that many of them are participating in scams and spam. 

The spam pages “used clickbait tactics and attempted to direct users to off-platform content farms and low-quality domains,” the researchers said. Scam pages tried to sell fake products or get people to divulge their personal information. 

The BBB offers a handful of tips to protect yourself from like-farming scams: 

  • Use your good judgment. If a post tugs at your heartstrings and isn’t about someone you know personally, be wary about its truthfulness. 
  • Don’t “like” every post in your feed. Scammers are counting on getting as many likes as possible, so be sure to only “like” posts and articles that are legitimate. 
  • Don’t give out your personal information, such as your full name, telephone number or address, to a person or company you don’t know. This includes commenting on a public post. 
  • Update your web browser. Make sure you always have the latest version of your browser. If you accidentally click on a scammer’s post, your browser will be more likely to warn you about a suspicious website.

The VERIFY team works to separate fact from fiction so that you can understand what is true and false. Please consider subscribing to our daily newsletter, text alerts and our YouTube channel. You can also follow us on Snapchat, Instagram, Facebook and TikTok. Learn More »

Follow Us

Want something VERIFIED?

Text: 202-410-8808

Source link

Written By

Leave a Reply

Your email address will not be published. Required fields are marked *