Artificial intelligence is almost everywhere these days, helping people get work done and also helping them write letters, create content, learn new things, and more. With it comes the fear of technology being too ubiquitous, as it could potentially replace some people’s jobs. AI must be used with caution, as it doesn’t necessarily provide the right information and can become biased, racist, or insulting. So, it’s important to use it smartly, knowing its shortcomings and potential flaws.
AI can be used in different ways, including conversational tools such asGoogle Bardand ChatGPT, but also in the form of solutions designed to create content, images, and even videos or soundtracks. Several services are available online, including Dall-E and Midjourney, which are open to the public and let anybody generate a fake image by entering what they’d like to see.
These are sometimes so powerful that it is hard to tell AI-generated images from actual pictures, such as the ones taken with some of thebest camera phones. There are some clues you’re able to look for to identify these and potentially avoid being tricked into thinking you’re looking at a real picture.
These tips help you look for signs indicating an image may be artificially generated, but they can’t confirm for sure whether it is or not. There are plenty of factors to take into account, and AI solutions are becoming more advanced, making it harder to spot if they’re fake. On the contrary, some artifacts do not necessarily mean an image is fake. It could indicate it’s been tampered with or edited.
Take a closer look at people
One of the first things you should pay attention to is how humans are represented in the picture. AI struggles to accurately reproduce human body parts because they’re complex, so paying close attention to these can help you identify if there’s something wrong with the image.
Look at the hands, feet, ears, and noses. Then assess whether they look unnatural or not. For instance, there may be inconsistencies, such as an unusual number of fingers, abnormal shape, or peculiar positioning. Similarly, look at facial details that might look strange, especially around the eyes and on the ears, as these are often harder to generate for AI. A tricky feature to reproduce for AI is wrinkles and lips, which need to be consistent across the face, and a discrepancy in these can also be a sign the image isn’t necessarily a real photo.
On the contrary, if a face looks too symmetrical or doesn’t have lighting reflections or natural imperfections, it could be an AI-generated one. The same applies to teeth, with too perfect and bright ones potentially being artificially generated.
Check all the details
Besides people’s bodies, it’s also important to look at all the elements in the picture, such as clothes and accessories. Check if these make sense or whether the shading and details are accurately represented. If there are animals or flowers, make sure their sizes and shape make sense, and check for elements that may appear too perfect, as these could also be fake.
Assess the overall scene
Besides the details, it’s important to pause and look at the big picture. First, check the lighting and the shadows, as AI often struggles with accurately representing these elements. Shadows should align with the light sources and match the shape of the objects casting them. Similarly, have a look at the background. If the pattern and textures are repetitive, it may be a fake picture.
Also, make sure the overall scene is realistic and use common sense. For example, it’s unlikely a lion would be having dinner with a family of penguins while wearing a pearl necklace and holding silverware with its paws.
Finally, if something feels awkward, fact-check unusual events online using a search engine, reliable sources, and news outlets. If you don’t find anything online or only data from unknown sources, the image may be AI-generated.
Investigate text and labels
If the image you’re looking at contains texts, such as panels, labels, ads, or billboards, take a closer look at them. AI-generated text can appear pixelated or stretched. Similarly, if there are logos, make sure they’re the real ones and aren’t altered. These signs can alert you to a potential AI-generated image.
Furthermore, if the content doesn’t make sense, is out of context, or contains weird phrases that a human is unlikely to write, the image you’re looking at is likely fake. This may seem obvious, but remember that these elements could be in the background of a deepfake image showing a celebrity visiting the North Pole, so scan for these minute details.
Leverage smart tools
If you have doubts about an image and the above tips don’t help you reach a conclusion, you can also try dedicated tools to have a second opinion. The first one is as simple as running a reverse image search on Google Images orTinEye.com, which will help you identify where the image comes from and if it’s widespread online. While they won’t necessarily tell you if the image is fake or not, you’ll be able to see if it’s widely available online and in what context.
Some AI-detection tools can do the work for you and assess whether a picture is authentic or AI-generated. These are relatively new and aren’t always reliable, but more options are showing up online to help you identify computer-generated images, such asDeepFake-o-meter.
Always use your best judgment
The above tips are merely guidelines to help you look at potentially abnormal elements in a picture. They can’t guarantee whether an image is AI-generated, authentic, or poorly edited. It’s always important to use your best judgment when seeing a picture, keeping in mind it could be a deepfake but also an authentic image.
Some tools have become particularly good at generating realistic images and may fool even the most detail-oriented people. However, most of them aren’t flawless and still leave tell-tale signs that the image isn’t natural, which should tip you off.
