Introduction

The rapid emergence of Generative Artificial Intelligence (GenAI) tools such as ChatGPT has already transformed the landscape of higher education (HE), offering learners resources to facilitate writing, coding, deepening subject understanding, and to produce AI-generated images. GenAI’s challenges and possibilities relating to learning and assessment have received considerable coverage (see for example QAA, 2023), but little is known about its potential impacts on mental wellbeing and social connectedness. In this blog post, we explore these impacts, drawing on an ongoing survey research project and our experiences as university-based educators. In doing so, we hope to build on recent field discussion shared in related BERA blog posts focusing on challenges posed by GenAI to higher education and considering GenAI’s possibilities from the perspective of graduate learning outcomes.

GenAI and social connectedness

Almost by definition, GenAI tool engagement involves computer interaction, but such engagement brings with it dialogue, and a collaboration with and between peers and educators. GenAI tools, used inventively by educators, may nurture social connectedness through challenging or promoting dialogue between learners to facilitate deeper learner–learner or educator–learner interactions than groupwork that follows traditional lines.

For example, a GenAI-based task completed by learners individually, built into the start of a session, could support exchanging ideas between peers in group work, then build outwards into a wider plenary. In this way, GenAI, included alongside other educational activities, may help learners engage with groupwork, and support interpersonal communication and mutual exchange of ideas.

‘GenAI, included alongside other educational activities, may help learners engage with groupwork, and support interpersonal communication and mutual exchange of ideas.’

GenAI tools may also hinder social connectedness within the learning experience. For example, excessive reliance on GenAI tools might unsettle the role of traditional, classroom-based interactions (such as peer discussion and collaboration), potentially hindering the development of supportive learning communities. A worst-case scenario might involve educators and learners enacting meaningless performances (such as those relating to classroom activities or assessment) where GenAI-enforced ‘empathy metrics’ begin to replace, govern or confuse higher-quality interactions and genuine educational collaborations (Selwyn, 2024).

An additional risk concerns potential dependence on GenAI tools in an educational environment increasingly saturated by online engagement (such as email, module Virtual Learning Environments). Such dependence might inhibit learner social connectedness given evidence showing links between internet addiction and social connectedness (Gawrych, 2022). Drawing on GenAI tools to successfully co-exist with traditional opportunities for social interaction between learners presents an immediate challenge for educators.

GenAI and mental wellbeing

There are well-established links between internet addiction and poor mental health outcomes (see for example Zewude et al., 2024), and assessment involving substantial engagement with GenAI tools could carry a mental health burden for learners.

GenAI is likely to be used simultaneously with other technologies and platforms including watching online guides (for instance on YouTube) and engaging in WhatsApp learner communities. The phenomenon of shifting between platforms and technologies has been noted in research with UK-based university students as ‘transferal between devices’ for its addictive potential (Conroy et al., 2023). We reflect that any assessment where GenAI is a requirement may carry resultant strains due to the multi-layering of technologies involved.

Assessments incorporating GenAI may exacerbate learner anxiety due to the ambiguity around the independent learning process and the possibility of being accused of academic misconduct. Relatedly, learners may assume that assessment integrating GenAI requires less time allocated to work planning, potentially leading to anxiety and distress when deadlines are not met. Finally, GenAI arguably marks the advent of a more involved, relational partnership and dialogue between learners and technology, bringing unpredictable and distinctive demands, strains and anxieties.

Conclusion

GenAI in HE presents both opportunities for enhancing learning experiences but also challenges to learner social connectedness and mental wellbeing. Future research must explore potential causal links between GenAI tool engagement, social connectedness and mental health. Navigating this new frontier, educators must strike a balance between leveraging GenAI’s advantages and safeguarding the emotional and social wellbeing of our learning communities to ensure that GenAI’s integration in education is beneficial, ethical, and truly transformative.

References

Conroy, D., Chadwick, D., Fullwood, C., & Lloyd, J. (2023). ‘You have to know how to live with it without getting to the addiction part’: British young adult experiences of smartphone overreliance and disconnectivity. Psychology of Popular Media, 12(4), 471–480. https://doi.org/10.1037/ppm0000425

Gawrych, M. (2022). Internet addiction in light of social connectedness and connectedness to nature. European Psychiatry, 65(S1), S596–S596. https://doi.org/10.1192/j.eurpsy.2022.1526

Quality Assurance Agency [QAA]. (2023). Reconsidering assessment for the ChatGPT era: QAA advice on developing sustainable assessment strategies. https://www.qaa.ac.uk/docs/qaa/members/reconsidering-assessment-for-the-chat-gpt-era.pdf?sfvrsn=38d3af81_1

Selwyn, N. (2024). On the limits of artificial intelligence (AI) in education. Nordic Journal of Pedagogy and Critique, 10, 3–14, https://doi.org/10.23865/ntpk.v10.6062

Zewude, G. T., Bereded, D. G., Abera, E., Tegegne, G., Goraw, S., & Segon, T. (2024). The impact of internet addiction on mental health: Exploring the mediating effects of positive psychological capital in university students. Adolescents, 4(2), 200–221. https://doi.org/10.3390/adolescents4020014