The bottom line: Artificial intelligence tools are being tested to help universities identify student distress earlier and offer support when counselling services are overstretched. Evidence suggests chatbots, emotion recognition and wearables can provide low cost, continuous monitoring and early intervention, but accuracy, fairness and privacy limits remain. Used carefully, these systems could strengthen mental health prevention, improve well-being outcomes and inform campus health policy without replacing human care.
University life can be an intense and isolating experience, with academic pressure, financial strain, and social change often colliding at once. For many students, anxiety, stress, and low mood become part of daily life, sometimes affecting sleep, physical health, and academic performance. As demand for mental health support continues to outstrip capacity on campuses, researchers are increasingly exploring whether artificial intelligence can help close the gap.
Recent research suggests that artificial intelligence tools may offer new ways to detect emotional distress early and provide support before problems escalate. By analysing patterns in language, behaviour, facial expressions, and physiological signals, these systems aim to identify subtle changes that are difficult to spot through traditional services alone. The goal is not to replace human care, but to supplement it with faster and more accessible support. The findings were published in Psychreg Journal of Psychology.
Dr Carolina Lunetti, professor at Guglielmo Marconi University in Italy, said: “The study is motivated by the growing mental health challenges faced by university students, who are increasingly exposed to academic pressure, stress, and emotional demands that negatively affect their psychological well-being, physical health, and academic performance.”
One area attracting particular attention is the use of chatbots designed to simulate elements of therapeutic dialogue. These digital tools can guide students through calming exercises, stress management techniques, and basic cognitive behavioural strategies. Studies indicate that many students are willing to engage with these systems, especially when support is available at any time without long waiting lists.
According to Lunetti: “Chatbot-based interventions, particularly those grounded in cognitive behavioural techniques, are generally well accepted among university students and have been shown to reduce symptoms of stress and anxiety while providing continuous, low-cost support.”
Another strand of research focuses on emotion recognition technologies. Artificial intelligence systems can now analyse facial expressions, voice tone, and written language to estimate emotional states with a high degree of accuracy. This approach may allow universities to monitor well-being trends across student populations and respond more quickly when risks emerge, particularly during high pressure periods such as exams.
Lunetti noted that “emotion-recognition technologies based on facial expressions, physiological signals, and linguistic analysis further enable a more objective and fine-grained assessment of students’ emotional states,” while also cautioning that “many existing tools rely on a single data source, focus primarily on basic emotions, and lack adequate validation within university populations.”
Wearable devices are also being explored as part of this broader picture. Smartwatches and similar technologies can track indicators linked to stress, including heart rate, sleep patterns, and physical tension. When combined with artificial intelligence, these data streams may help identify early warning signs of emotional overload and prompt timely interventions.
Despite this promise, researchers caution that current systems have important limitations. Many tools rely on a single source of information rather than integrating data from multiple channels, which can reduce accuracy. There are also concerns about how well emotion recognition systems perform across different cultural and demographic groups, raising questions about fairness and reliability.
Expanding on these challenges, Lunetti said that “these shortcomings limit the accuracy and clinical applicability of current AI-based mental health systems,” underlining the need for more robust and inclusive designs.
Privacy and accessibility remain central challenges as well. Continuous monitoring of emotional states requires careful handling of sensitive personal data, along with clear safeguards around consent and data use. At the same time, tools must be designed so they are easy to access and do not widen existing inequalities in mental health support.
Looking ahead, researchers argue that the most effective approaches will combine daily self-reporting with objective indicators from wearables and emotion recognition systems. This integrated model could offer personalised, preventative support that complements existing university services rather than competing with them. Lunetti echoed this view, stating, “Future research should focus on developing integrated AI systems that combine ecological momentary assessment, wearable sensors, and advanced emotion-recognition algorithms to monitor students’ emotional states in real time.”
As pressure on student mental health continues to grow, artificial intelligence is increasingly being seen as one part of a wider solution rather than a cure in itself.
