GenAI supervision can create a more ‘humanised’ experience

Rapid advances in generative artificial intelligence (GenAI) have made it increasingly important in graduate research settings, serving as a valuable tool for enhancing research experiences. Its adoption has stimulated sustained debates on AI’s capacity to augment – rather than replace – human expertise in academic mentorship.

Scholars and practitioners now embrace a complementary perspective that synthesises the unique strengths of both human supervisors and GenAI to enrich postgraduate students’ research experiences.

The evolving dialogue surrounding GenAI in educational contexts represents a complex, dynamically unfolding landscape, which demands continued scholarly investigation to comprehend supervisor-GenAI collaboration.

However, empirical evidence remains scarce for sustained supervisory-GenAI partnerships that systematically address postgraduate students’ evolving research needs through coordinated task allocation and adaptive feedback mechanisms.

We propose that analysing the relationship between supervisor support, GenAI support and postgraduate students’ research experiences – particularly in terms of generic skills, research skills and research satisfaction – represents a critical foundation for understanding this emerging dynamic.

What are the interrelationships between them? How do supervisor support and GenAI support influence the research experiences of postgraduate students? To what extent can GenAI potentially complement the unique contributions of supervisors in shaping research experiences?

To answer these questions, we employed an explanatory sequential mixed-methods design.

We first involved an online questionnaire survey of 1,515 postgraduate students from Double First-Class universities in mainland China, comprising 811 MPhil students and 704 doctoral students, assessing how supervisor support and GenAI support influence their research experiences.

Next, we examined how supervisor support and GenAI support function synergistically and to what extent GenAI support complements supervisor support through semi-structured interviews with 20 postgraduate students.

Complementarity

We found that supervisor support and GenAI support operate as independent yet complementary mechanisms, each contributing unique values to postgraduate students’ research experiences. Optimal outcomes emerged from their collaboration rather than from either support working in isolation.

While GenAI support was a relatively weaker predictor than supervisor support – particularly in terms of skill development and research satisfaction – the findings underscored the necessity of maintaining both forms of support.

Most crucially, even when GenAI provided substantial support, inadequate supervisor support remained a key contributor to postgraduate students’ negative research experiences.

This demonstrated that GenAI neither replaced nor diminished supervisors’ irreplaceable role in shaping postgraduate students’ research experiences. Rather, it accentuated their unique contributions.

With respect to the ‘how’ of influencing, we adopted semi-structured interviews to elicit postgraduate students’ perspectives. We mapped a support ecosystem in which human supervisors and GenAI fulfilled distinct yet complementary functions in terms of academic and emotional support.

Supervisors provided comprehensive academic support by serving as academic guides, offering critical research resources, building academic networks and cultivating research values, while simultaneously fostering high-quality relationships that offered crucial emotional support.

In parallel, GenAI assisted in non-core research tasks and offered out-of-the-box ideas, alongside acting as a timely psychological companion.

This dual-support system reduced postgraduate students’ over-dependence on supervisors and addressed their unmet needs, particularly bridging gaps when postgraduate students hesitated to approach supervisors about basic or non-academic concerns.

The personalised, task-specific nature of GenAI support proved particularly valuable within China’s performance-driven academic culture, where supervisor support was increasingly strained by the ongoing expansion of postgraduate education and its accompanying challenges to maintaining quality mentoring relationships.

Three complementary modes of collaboration

The influence exerted by supervisors and GenAI on the research experiences of postgraduate students was discernibly distinct.

The divergent yet interconnected nature of these influences prompted us to examine their synergistic potential, resulting in the identification of three complementary modes of supervisor-GenAI collaboration.

First, GenAI served as a preliminary intellectual scaffold – handling routine analytical tasks and generating initial insights – while supervisors leveraged their scientific intuition and extensive knowledge networks to transform these foundations into meaningful academic contributions through their seasoned academic judgement.

Second, GenAI functioned as a bounded technical resource for specific research tasks while supervisors served as irreplaceable academic mentors, shaping postgraduate students’ scholarly development and research identity.

Third, GenAI provided constant availability and passive listening, complementing supervisors’ deeper, experience-based emotional support to address postgraduate students’ immediate psychological needs and long-term academic development.

A more ‘enriching and humanised’ experience

We envision a transformed university environment where the integration of GenAI with supervisors’ academic expertise creates more enriching and humanised research experiences for postgraduate students.

In this framework, GenAI serves as a foundational academic scaffold, while supervisors employ their academic intuition and experiential wisdom to transform technical inputs into meaningful contributions and offer profound guidance, collectively nurturing postgraduate students’ immediate psychological resilience and long-term academic development.

This relationship transcends mere functional collaboration – supervisors emerge as irreplaceable academic guides whose influence extends beyond technical oversight to shape postgraduate students’ research identity and professional development.

This integrative approach strategically amplifies supervisors’ leadership potential while preserving the essential human elements of supervisor support.

Overall, we underscore the value of GenAI as a potential solution to systemic challenges in postgraduate education, paving the way for exploring how GenAI interventions can be leveraged to address structural issues in higher education systems.

To bring this vision to life, higher education institutions should establish lasting partnerships with leading AI enterprises to co-develop customised GenAI tools for addressing postgraduate students’ research needs.

Concurrently, they need to implement ‘GenAI Support’ projects that strengthen traditional advising through AI-powered feedback and review systems.

Supervisors should be encouraged to develop prompt expertise about AI and create detailed instructional guides to enhance postgraduate students’ effective utilisation of GenAI.

Postgraduate students are urged to acquire proficiency in designing and employing well-crafted prompts to execute diverse research tasks, while systematically evaluating GenAI tools’ efficacy to select the most suitable tool for their individual research needs.

Yating Huang is currently an associate professor, working in the College of Education at Zhejiang University in China. Her publications appear in a wide range of academic journals, such as Higher Education, Studies in Higher Education, Higher Education Research and Development, Compare, Studies in Educational Evaluation, Educational Studies and Professional Development in Education and System. Sihui Li and Zihan Liu are doctoral students under Huang’s supervision at the College of Education, Zhejiang University.

This article draws on their study, ‘Can GenAI complement supervisor support in shaping postgraduates’ research experiences? A mixed-methods approach’, which has just been published in the journal Studies in Higher Education.

This article is a commentary. Commentary articles are the opinion of the author and do not necessarily reflect the views of University World News.