Six telltale signs an image is AI generated

Instances of artificially generated imagery are growing rapidly online, especially on social media. While the use of these images isn’t always bad – hey, who doesn’t love pictures of giraffes dressed as clowns eating ice-cream?! – it’s the intention behind it that matters.

Some fraudsters are using image generation tools to create fake job ads, service offerings, and products, to trick you into believing they are a legitimate business, so you share personal information or your card details. There’s even been cases of AI generated images used in complex romance scams. 

  

What is artificial intelligence (AI) generated imagery?

AI generated imagery refers to images or artwork created by AI software in response to a prompt, usually text based. Using algorithms and machine learning, this tool can produce high-quality and realistic imagery, from crisp photographs to intricate artworks.

Here are six things to ask yourself to help identify fake images the next time you’re scrolling your social feed. We'll use the above AI generated image as an example.

1. Are they anatomically correct?

It’s not uncommon for AI images to feature hands with extra fingers, one too many teeth, or a bonus arm. While this was a lot more common in the early days of AI images hitting the social sphere, it is still happening; particularly in scenes featuring multiple ‘people’. It’s worth paying close attention to any people, particularly those who appear in the background of the image, as less attention is paid to their finer details.

2. Is the text readable?

From jumbled letters and numbers on a street sign to blurry and distorted text in a brand’s logo, AI image generators can struggle to recreate text in visuals. At first glance everything may seem in order, however, take a moment to check if the text is readable, free from errors, and appears correctly.

3. Does it look heavily filtered?

Images produced by AI can often appear overly glossy, unnaturally smooth, and contain inconsistent lighting, often mistaken for a beauty lens or image filter. If an individual or business you come across uses this type of imagery online, approach with curiosity and do some independent research into the brand and/or profile – who doesn’t love a deep dive?

4. Does it make sense?

Often the closer you look at an AI generated image, the more it doesn’t make sense. Is there an unidentifiable object with no clear purpose in the background? Maybe there is an unexpected doorway or a path leading to nowhere. Once you see the error, you won’t be able to un-see it.

5. How does it make you feel?

Manipulated images are often designed to evoke strong emotions, such as greed, jealousy, anger, fear, or outrage. This can cloud your judgment and make you more likely to believe the image without questioning it.

6. Is the image from a trusted source?

In some cases, the image source or the context surrounding the image are not at all aligned. It’s good practice to check if the source is reputable and the context is relevant.

These signs alone aren’t always deal breakers when it comes to separating real from fake. Like us humans, AI is always learning, which means these signs are likely already being corrected as you read this.

This doesn’t mean you have to stop using the internet or social media. The digital world has become a big part of our lives, which is why it’s so important to navigate social media and AI with curiosity and plenty of caution.

Next time an interesting image catches your eye on social media, try the following before clicking any links or sharing it:

• Ask yourself the above questions and look at the image carefully.
• Think before you click a link or provide personal details to anyone.
• Do your own independent research into the offer, person, or business.
• Phone a friend or family member and talk it through with them.