Can we tell the difference between an MI image and a photo taken by a photographer?

Artificial intelligence (AI) is playing an increasingly important role in our daily lives. We encounter it at every turn - on websites, in texting, in chatting, in smart homes and in generating pictures. The MI algorithm is as close as possible to the human brain, but it cannot replace it completely. Therefore, if we have well-developed critical thinking, we will be able to distinguish between an image created by a real person and an image created by MI. Here are 5 tips on how to detect the presence of MI in photographs. 

Each photo has an author

Just that simple. When you look at a photo, check the metadata and make sure there's an attribution, as this is mandatory when publishing an image. If you can't find the author, or there's a special watermark in the margin, there's a good chance that MI has put their finger on it. The fact that MI was used in the creation of the photo is sometimes also indicated in the comments.

Asymmetry in detail

A closer look at, for example, people's faces or other parts of the body can reveal unnatural proportions or asymmetries in details, such as different earrings, strange teeth, odd hands, fingers. Particular attention should be paid to the pupils of the eyes, as they are not yet really reproduced in MI - wrong shape, different light reflection. But at first glance everything may seem fine. 

Generating portraits of people is a slippery business. It marks a dangerous trend for the future, with the risk of identity theft and the creation of fake profiles. But that is a job for other bodies.

Unnatural background

In photographs, the background may look like a blurred texture or appear artificially painted. The more unusual and uncommon colours you notice, the more likely it is that MI is at work. Anomalies in detail, shading, reflections also signal unnaturalness and MI fantasy, because physics is its weak point. The inconsistencies can also be seen in the various shapes - straight lines are shown as curved, while surfaces that in real life have a certain texture look smooth.

Look at the photo longer and more carefully

The reason is quite trivial. It turns out that the average person scrolls through a screen for a very brief moment - about 13 milliseconds - so at speed, it's easy to miss the fakeness of a photo if a blunder doesn't catch your eye right away. So next time you see a particularly interesting or funny picture, hang on to it a little longer - it might not be real.

Use tools that recognise photos made in MI

Just as artificial intelligence has been created, so too have the means to recognise it. These analysis tools or applications can try to understand whether a photograph is a product of humans or AI. Like a text editor, but for photographs. In this case, an advanced algorithm that works like a microscope is used to decipher the masterpieces of the photoshop. Since every image has the same error rate, if you notice strong differences in the same image, there is a risk that the photograph is artificial.

Any innovation that makes our daily lives easier is welcome because it opens up new possibilities. The problems start when it begins to threaten our security, integrity and privacy. It is to be hoped that artificial intelligence images will not be able to completely replace human work, because a photograph is not just an image, it is also an emotion conveyed by the photographer to others. A computer, unfortunately, cannot feel this. So be vigilant and don't be fooled!