Media Union: Are we defenseless against artificial intelligence?

According to European surveys, 96 percent of 13-18 year olds have used artificial intelligence in the past year (40 percent on a daily basis), and 51 percent of teenage boys also have emotional or “romantic” conversations with these tools. The situation is no different among adults: 30 percent of workers in the EU now use AI in their work. A recent experiment by the Media Union Foundation and Cyber ​​Shield showed that despite the high rates, it is far from certain that we would be better at recognizing what is real and what is artificial. In fact, according to the latest financial analyses, one in two financial fraud attempts now involves AI in some way…

Young, old – and AI

We showed a pair of twin boys and their grandmother, as well as their sister two years older and a man of grandfather’s age the same images and videos, and they all had to decide whether they were seeing real footage or content generated by artificial intelligence. The experiment did not include any deceptive or emotionally charged “grandchild” videos: ordinary-looking scenes, landscapes, and portraits were shown on the screen. There were some that seemed realistic at first glance, but actually hid absurd details – for example, a “cauliflower plantation” in Badacsony – and there were recordings in which every pixel was “flawless”.

The young people confidently looked for typical signs of AI: strange textures, disproportionate colors, protruding shadows, “video game” skin, small slips between mouth movements and voice. The older people, on the other hand, looked at how lifelike what they saw was: whether they could imagine the location, whether people were “like this” in real life. Although the contrast was spectacular, mistakes were not uncommon in either generation.

In connection with the experiment, deepfake and AI researcher Marcell Mezriczky analyzed in detail how AI changes our perception of reality. As he said, young people seem to have an advantage, as they have grown up using AI tools: it is natural for them that images and videos can be filtered, “tuned”, or even completely synthetic. “However, this routine can easily turn into overconfidence. Young people are less critical of sources, they trust themselves, and do not assume that digital modification can be so lifelike that they don’t even notice it.”

The older generation makes mistakes differently, they primarily rely on emotional conviction: if the character’s voice, gaze, and gestures seem familiar, they are more likely to accept the content as real. This logic often works well in everyday life, but in the age of AI, it is precisely this that becomes a risk. “The older generation believes content based on emotional conviction. This makes them vulnerable to AI-powered grandchild scams and manipulated calls.” It is no coincidence that 56 percent of financial professionals consider social engineering, or psychological manipulation, to be one of the most powerful AI-supported fraud techniques today.