People are worse than expected at distinguishing real photos from AI-generated ones.

People have a harder time than expected distinguishing between who in a photo is a real person and who is an AI-generated image.

.in_text_content_22 { width: 300px; height: 600px; } @media(min-width: 600px) { .in_text_content_22 { width: 580px; height: 400px; } }

This was established by scientists from the University of Vterloo (Canada) after conducting a corresponding study, reports EurekAlert.

As part of the study, 260 participants were given 20 unsigned photographs:

• 10 of them were images of real people obtained from Google searches;
• The remaining 10 were created using two popular artificial intelligence image generating programs, Stable Diffusion or DALL-E.

Participants were asked to label each image as real or AI-generated and explain why they thought so. As a result, only 61% of participants were able to distinguish people generated by artificial intelligence from real ones, which is significantly lower than the 85% that the researchers expected.

People are not as skilled at discrimination as they think, the researchers note.

It is noted that study participants paid attention to details such as fingers, teeth and eyes as possible indicators when searching for content created by artificial intelligence, but their assessments were not always correct.

The researchers note that the nature of the study allowed participants to look at the photographs for a long time, while most Internet users look at the image in passing.

People simply flipping pages or short of time will not notice these clues, the researchers note.

Scientists note that the extremely rapid pace of development of AI technology makes it difficult to understand the potential for malicious or nefarious actions that images generated in this way carry. In addition, the pace of academic research and legislative developments often lags, so since the study began in late 2022, AI-generated images have become even more realistic.

Disinformation is not new, but the tools of disinformation are constantly changing and evolving, the scientists note. – It may come to the point that people, no matter how prepared they are, will find it difficult to distinguish real images from fakes. That’s why we need to develop tools to detect and counter it. This is akin to a new arms race with artificial intelligence.

Source: EurekAlert

Source: Racurs

David

I am David Wyatt, a professional writer and journalist for Buna Times. I specialize in the world section of news coverage, where I bring to light stories and issues that affect us globally. As a graduate of Journalism, I have always had the passion to spread knowledge through writing.