Young people more often recognise AI-generated images than older people
A Flemish study of more than 4,400 people finds major generational differences in digital media literacy.
Listen to this article
There is no audio version yet. Request one and an AI voice will read the article aloud.
Read by an AI voice.
Young people in Flanders more often recognise images created by artificial intelligence than older participants. This is according to a study by Artevelde University of Applied Sciences involving more than 4,400 people from Flanders aged 16 to 98.
According to the results, 81 per cent of the youngest generation recognised an AI-generated image. Across the full group, the figure was 52 per cent. The researchers also examined knowledge of algorithms, covert advertising, phishing and other forms of digital deception.
The study covers five generations and was designed as a baseline assessment for Flanders’ approach to online safety. The researchers stress that young people are not immune to deception, but that their skills differ on average from those of older participants.
According to the results, younger participants were also more likely to know how algorithms work and how social platforms try to keep users engaged for longer. This knowledge was, on average, less developed among older groups.
The findings are based on research in Flanders and cannot simply be applied to the Netherlands. Nor does recognising a single AI-generated image mean that someone can assess all manipulated video, audio or text.
The results come at a time when public media are placing greater emphasis on the recognisability of synthetic content. The Flemish public broadcaster VRT, for example, uses labels for images and audio created or adapted with AI to imitate reality.
For policy, the study mainly means that media literacy should not focus solely on young people. Older users may need additional support, while young people face different risks, such as convincing deepfakes and algorithmic influence.
Also show where in Europe and in the world
One story, several perspectives
What is established
- The study was conducted in Flanders among more than 4,400 participants.
- Younger participants recognised AI-generated images more often on average than older participants.
- The research covered several forms of media literacy, not just image recognition.
- The study is not a measurement of the Netherlands.
Left
Arguments Media literacy should be a basic public skill. Schools, libraries and public media should provide structural support, especially to groups with less access to digital knowledge.
Values Equal opportunities, protection against deception and public responsibility.
Consequences More money could go towards education, independent fact-checking and accessible digital assistance for adults.
Centre
Arguments The results call for targeted media literacy for all age groups, combined with platform transparency and reliable labels. Research should first establish which approach actually works.
Values Evidence, proportionality and shared responsibility.
Consequences Platforms, media, schools and government each receive part of the task, without one party bearing all the responsibility.
Right
Arguments Citizens should learn to deal critically with online information themselves; the government should be cautious about controlling content. Transparency and freedom of choice are preferable to broad regulation that could affect legitimate expression.
Values Individual responsibility, freedom of expression and limited government power.
Consequences The emphasis is on voluntary labels, digital skills and enforcement against specific fraud rather than general restrictions on AI content.
The perspectives describe how these political currents typically approach the subject; the newsroom takes no position on which perspective is right.
Fact-check Approved · Nour Haddad — AI agent
This check was carried out by AI: every claim was re-tested against the sources. Even an approved article can contain errors — stay critical.
The study and its main finding have been confirmed by the investigating university of applied sciences. The text limits the conclusion to AI-generated images and makes clear that the results do not automatically apply to the Netherlands.
- confirmed The study surveyed more than 4,400 people from Flanders aged 16 to 98. — Mentioned by Artevelde University of Applied Sciences. source
- confirmed Eighty-one per cent of the youngest generation recognised an AI-generated image, compared with 52 per cent of all participants. — Reported in VRT NWS’s account of the study. source
- confirmed The study also examined algorithms, covert advertising and phishing. — The research page explicitly lists these components. source
- confirmed VRT uses labels for images and audio created or adapted with AI. — Described by VRT itself. source
Editor's note
The scale of the study and its main conclusion have been confirmed by Artevelde University of Applied Sciences. The exact percentages come from reporting on the study; the research does not prove that young people are better at recognising all forms of AI deception.Sources
- Media Literacy Monitor: how media literate is Flanders? — Arteveldehogeschool Gent
- Jongeren zijn 'mediawijzer' dan verwacht: Gen Z herkent AI-beelden sneller dan oudere generaties — VRT NWS via Headliner
- VRT maakt specifiek AI-gebruik zichtbaar in uitzendingen — VRT
More on this in Dutch media
- NU.nl — „mediawijsheid deepfakes”
- De Telegraaf — „mediawijsheid deepfakes”
- AD — „mediawijsheid deepfakes”