Young adults in the Netherlands aged 18 to 25 correctly identify AI-generated human faces only 47 percent of the time, according to a recent analysis by software company Floowy.ai and digital marketing agency DGTLbase.
The findings challenge the common assumption that younger generations, having grown up with smartphones and social media, are naturally better equipped to detect digital deception online.
In practice, the high volume of content young people scroll through daily often leads to quick, passive viewing rather than detailed scrutiny, making synthetic images harder to spot.
Misplaced confidence in digital skills
The study highlights a significant gap between how well young adults think they can spot fake images and how they perform in real tests.
Survey data shows that only 30 percent of Gen Z respondents admit to having difficulty recognizing AI content, compared to 65 percent of older generations.
However, their actual performance rate of 47 percent is barely better than random guessing and shows no major advantage over older age groups.
Artificial faces rated as more trustworthy
The Dutch analysis aligns with global findings from institutions like the World Economic Forum, which have noted that computer-generated faces are often perceived as more realistic than real photos.
In the Dutch tests, participants rated AI-generated faces as 8.3 percent more trustworthy on average than genuine photographs of actual people.
This heightened perception of trust presents clear risks for online fraud, phishing attempts, and misleading commercial campaigns target audience members in major Dutch cities like Amsterdam.
Upcoming European regulations
Starting in August, new regulations under the European Union’s landmark AI Act will require companies to clearly label synthetic images, videos, and text across the region.
The goal is to ensure consumers across EU member states know immediately when digital media has been generated by artificial intelligence.
Researchers involved in the study warn that labeling alone may not solve the problem if young users remain overly confident in their ability to filter out fake images without help.
