In March 2026, a statistical study was published on the impact of the use of artificial intelligence (A.I.) on the mental health of young Europeans. This survey was carried out in January 2026 by the CNIL (Commission nationale de l’informatique et des libertés, the French data protection authority) and VYV (Union mutualiste de groupe) on a sample of 3,800 young people aged 11 to 25. The young people surveyed included 1,000 French, 1,000 Swedes, 1,000 Germans and 800 Irish, from several social categories. The survey was conducted through an online Ipsos questionnaire.
This study paints a picture of the different ways young people use A.I., through the prism of their mental health.
First of all, 86% of the young people surveyed use A.I. For 71% of them, the highest trend, it is a writing tool, to assist with their work or studies. Generative A.I. for images, music or text is used by 42% of respondents as a leisure activity. As for conversational A.I.s, 37% of the French respondents say they use them regularly, and 35% use these A.I.s to advise them when stressed in their professional or personal life. As for frequency of use, 25% of young people say they use A.I. daily, and 58% at least once a week.
As the subject of the survey is the use of A.I. tools and mental health, it therefore asks about the well-being felt by young people at the time of the questionnaire. To quantify this, the study uses the GAD-7 method, a scale measuring the prevalence of generalised anxiety disorder.
Thus, one young person in three in Ireland and Sweden, and one in four in France, say they do not feel well overall. For 20-25-year-olds in France, it is two in five. In light of this first finding, A.I. appears as the third easiest person for young people to contact in times of distress, after friends and parents. This ease of contact tends to increase for young people experiencing significant distress, some of whom receive psychological or psychiatric follow-up. This trend, seen in France, is confirmed in the other countries surveyed.
Among respondents in France, almost one young person in two (48%) says they turn to A.I. as a personal adviser. This statistic is higher in other European countries, notably Germany, where 63% of young people said so. Several reasons are given to understand what interest these young people have in this “A.I. adviser”. First, it is a tool available 24 hours a day, 7 days a week. Second, several respondents say it is easier to communicate with an A.I. than with those close to them on certain subjects, an argument that is more prevalent among young people in significant distress. Finally, one of the arguments put forward by 17% of people is that consultations with psychologists would be too expensive for them.
Consequently, three young people in four use A.I. tools to discuss their problems, even going so far as to consider A.I. as a life adviser (64%), as a psychologist (46%), or even as a lover (20%). We see here that, for most of the French respondents, A.I. is anthropomorphised.
Despite the statistics given above, the panel is evenly divided on whether A.I. is an effective tool for feeling better, and 34% of them think A.I. gives better advice than humans. Once again, the figures on the beneficial effects of conversational A.I.s as a way of feeling better are higher among young people experiencing significant distress.
With this massive use of A.I. tools in everyday life, the question of awareness of the risks of sharing personal information with A.I. tools emerges. 60% of the panel say they are well informed about the economic functioning of A.I., but 68% of respondents admit they do not know what becomes of the information and data they share with A.I.
Moreover, one in two young people surveyed is aware that A.I. companies are not guaranteed to respect confidentiality. In addition, the panel is divided on the balance of benefits and risks of using A.I. A first third of respondents think that talking about their personal life with A.I. carries more risks than advantages, a second third think the opposite, and the last third consider risks and advantages to be equivalent. This awareness of risks goes beyond the question of confidentiality, and one young person in two considers that conversational A.I.s can lead to isolation and a drop in overall well-being.
Finally, 85% of young people express a strong demand to be better informed about how information shared with artificial intelligence tools is handled. A fortiori, these questions could extend to the business model of A.I.s (notably free ones).
To conclude, the study sheds light on the place artificial intelligence tools take in the daily lives of the young people surveyed. We note that their main use is as a professional or school assistant. Added to this is a strong correlation between the use of conversational A.I.s for personal life and the person’s level of distress. This raises the question of a possible social and/or family isolation that would lead to this specific use of A.I., to be correlated with other psychosocial and economic factors (social and family background, available resources, place of residence, etc.).
For more information, you will find the detailed study here:
https://www.cnil.fr/sites/default/files/2026-05/ipsos_bva_les_jeunes_et_l_ia.pdf
N.L.
