A digital training programme designed to enhance empathy and reduce cyberbullying among teenagers has failed to deliver the intended results, according to a newly published international study. The findings were published in the International Journal of Bullying Prevention.
The initiative, tested on 332 secondary school students aged 11 to 17 across Brazil, Germany, Italy and Spain, used a virtual learning companion in a simulated social media setting. This companion, depicted as a robotic dog named Tutxi, guided users through interactive lessons on empathy, including a video segment and exercises aimed at triggering emotional responses to online bullying scenarios.
Johanna Börsting, PhD, a researcher from Ruhr West University of Applied Sciences, said: “Social media can be a challenge, especially for children and young people, for example when they are confronted with cyberbullying. As parental supervision or external control is neither realistic nor desirable, we have developed and evaluated a prototype of a virtual learning companion to support young users on social media. This virtual learning companion comments on problematic content and raises awareness of the negative consequences of cyberbullying. It is particularly important to us to take users’ autonomy into account, expand their skills and digital capabilities and strengthen their empathy. Overall, we want to strengthen users’ self-protection skills and sensitise them to a self-confident and critical approach to negative content in social media with the help of virtual companions.”
Participants were randomly assigned to either an experimental group, which received empathy-focused content, or a control group, which engaged with a general social media awareness video. Both groups were then shown a meme portraying a cyberbullying incident and asked to reflect on it.
Contrary to expectations, there was no significant reduction in cyberbullying intentions nor a measurable increase in empathy among those who completed the training. Both cognitive empathy, which refers to understanding another person’s emotions, and affective empathy, which involves sharing in those emotions, remained largely unchanged after the intervention. Furthermore, the intention to help a victim or avoid reinforcing a perpetrator did not differ between the groups.
Veronica Schwarze, a researcher from Ruhr West University of Applied Sciences, clarified: “Unfortunately, we did not observe a reduction in cyberbullying intentions or an increase in empathy after the empathy training with the virtual learning companion. However, we did find some cross-national differences in users’ cyberbullying intentions and their cognitive and affective empathy. This offers interesting insights for context-sensitive intervention programmes.”
The data revealed distinct regional differences. Participants from Italy demonstrated higher levels of both cognitive and affective empathy than their German counterparts. Meanwhile, bullying intentions were significantly higher among teenagers in Spain and Brazil compared to those in Germany. These findings suggest that national and cultural contexts may influence how young people perceive and respond to online abuse.
Despite its shortcomings, the study’s developers maintain that the digital format holds promise, especially given the widespread use of smartphones and social media among young people. The empathy training, delivered via a familiar chat-style interface, was designed to mimic platforms like WhatsApp. Most participants responded positively to the format, with nearly half describing the empathy video as helpful, even if it did not translate into measurable behavioural change.
The researchers acknowledged that a single-session intervention may not be sufficient to produce lasting effects. Dr. Johanna Börsting outlined future directions by noting, “Our studies provide six important insights for the future. First, empathy trainings should be repeated and ideally conducted in real time rather than in a single session. Second, various real-life examples with positive outcomes for model-based learning should be considered. Third, future research should focus on users’ perceived control and influence over the situation, which is important for their autonomy. Fifth, as we found cross-national differences, different patterns and conditions in different settings need to be considered. Sixth, instead of developing multiple interventions for different contexts, we propose a sustainable approach by developing a basic empathy training that can be extended with specific tools so that it is possible to adapt the intervention to different social and cultural contexts.”
Additionally, the study highlighted the need to address adolescents’ sense of control in bullying situations. Simply raising awareness is unlikely to be effective if teenagers do not feel empowered to act. Developing interventions that align with young people’s attitudes and perceived capabilities may be key to shifting their behaviour.
The use of a virtual learning companion was seen as a non-threatening way to introduce sensitive topics. Unlike a teacher or authority figure, the companion functioned more like a peer, guiding users through questions and content in a conversational tone. While this approach may encourage engagement, the study authors note that deeper behavioural change likely requires a combination of digital and human-led support.
