The world we live in today is highly technological and advanced. Communication is essential, just as staying constantly connected is. For a long time, people have tried to fight disinformation and fake news, but it is often easier to believe a story when it is accompanied by photos or videos, because these elements seem to provide concrete proof.In recent years, however, artificial intelligence has made this task even more complex: it is now possible to create fake images and videos that look real. But how does it work?
For images, software like MidJourney is used, which is known for creating artistic yet realistic content, and Stable Diffusion, an open-source and highly customizable program.When it comes to videos and deepfakes, there are platforms like Synthesia and Reface, which allow the creation of realistic videos with synthetic faces or even voice imitation.So, how can we recognize an image or video generated by artificial intelligence? There is no foolproof method, but it is possible to notice some signs: wrong proportions, details that look too perfect or unnatural, and incorrect or distorted texts and logos. Also, through advanced forensic analysis, such as examining digital noise patterns or inconsistencies in lighting, it is possible to tell a fake from a real piece of content.Today, recognizing AI-generated content is a complex challenge. However, by combining technological tools, critical thinking, and forensic techniques, it is possible to significantly reduce the risk of being misled.
Author: Ms Maria Lagani, psychology graduate- Master’s student. Research Team for JUMP staff.