Three-quarters of middle and high school students use artificial intelligence daily, and a quarter of children and adolescents who are victims of cyberbullying turn to tools like ChatGPT, Claude, or Gemini to seek help. This is shown by a survey from the e-Enfance association, conducted in partnership with Caisse d’Épargne. However, this support remains limited, according to the association’s experts.
According to this study, the proportion of middle and high school students who use artificial intelligence daily has increased by ten points in one year, reaching three-quarters. The e-Enfance association notably manages the national helpline 3018 against digital violence. The survey interviewed 1,612 parent/child pairs from the same family.
Artificial intelligence is not just a refuge. According to the barometer, one in three victims says they have been targeted by humiliating content generated by AI. It can thus become a weapon in the hands of cyberbullies, while also being a recourse for other young people facing their bully.
An interlocutor who does not react like an adult
For Samuel Comblez, deputy general director of the association, if young people turn to AI, it’s because they feel they keep control of the conversation. ‘The advantage when addressing AI is that there isn’t the reaction, or even the overreaction, of an adult who might be very angry, very worried, or frightened by their child’s revelation. You can also control the pace of the exchange, you can pause, you can come back, and you also have, I would say, control over what will happen after the facts are revealed,’ he explains.
With parents, the reaction is quite different. ‘Obviously, we will act immediately. And what victims often say is that they feel dispossessed of their own words. The advantage of AI is that it won’t go further than a few pieces of advice,’ continues Samuel Comblez. According to the victims surveyed, in 80% of cases, AI helped them reduce or resolve their problems.
Very real limits
The responses provided by AI can take several forms: advice on how to react to the person bullying, a helpline number, or simply comfort. However, several studies, particularly in the United States, have shown that chatbots are not always good advisors.
Above all, these tools cannot detect all distress signals. ‘Artificial intelligence, however, does not hear silences. It does not see facial expressions when someone is not doing well. So it misses a lot of important information to be able to help the victim in an appropriate way,’ emphasizes Samuel Comblez.
The e-Enfance barometer also sends a message to adults. Only half of the bullied young people who turned to AI also confided in their parents. For Samuel Comblez, this should encourage adults to listen better to young people.
