Generative AI (GAI) has sparked debate among academics. Some see its potential to boost productivity and simplify tasks, while others worry it could undermine rigorous scholarship. Currently, most discussions focus on student use and how educators should monitor it in assessments (Cotton et al., 2023; Xia et al., 2024). However, there’s also concern about how GAI might affect academics and the nature of original research.

Application of GAI for research

Many are familiar with the ability of GAI to generate, summarise and synthesise text, but its research capabilities are expanding. Recently, OpenAI and Google released their Deep Research features, designed to autonomously search the internet and then synthesise the findings. Such tools could allow researchers to conduct research in minutes rather than weeks.

Another intriguing advancement is AI ‘reasoning models’, which handle complex analytical tasks previously considered uniquely human. These models can take a research problem, analyse patterns from large datasets and refine their responses step-by-step, as a human researcher might think through a complex issue. However, there are caveats, and the OpenAI project, Humanity’s Last Exam (HLE), is attempting to benchmark AI models – and at the time of writing, there is still ‘a significant gap between current capabilities and expert-level academic performance’ (Phan et al., 2025). Nevertheless, GAI tools are continuously evolving, and these features give us a hint of their direction.

Role of positionality within GAI-supported research

To improve academic research, ethical considerations must be part of both the use and design of GAI. Yan et al. (2024) suggest a human-centred approach to AI development, adding checks and balances to ensure tools are fair and inclusive. This approach is similar to the moderation and validation processes used in higher education and research peer-reviews.

‘A critical issue is that GAI relies mostly on Western-centric datasets and algorithms, so may embed and perpetuate certain knowledge structures and assumptions.’

A critical issue is that GAI relies mostly on Western-centric datasets and algorithms, so may embed and perpetuate certain knowledge structures and assumptions. This poses an ethical problem, as GAI may inadvertently reinforce existing biases, raising issues related to fairness, inclusivity and global representation in research. While human researchers should be transparent about their decision-making processes, AI systems may struggle to recognise or articulate their inherent biases. For example, when preparing this article, we asked a GAI tool to produce a report and then declare its positionality, which it initially said was not possible. However, when prompted further it acknowledged its built-in bias, positivist leanings, and reluctance to challenge the status quo.

If GAI is to enhance – rather than compromise – academic research, ethical considerations must be embedded not only in its use but also in its design. Yan et al.’s (2024) proposed human-centred approach to AI development presents an opportunity for educational researchers to critically evaluate research ethics, shifting from a narrow compliance focus to one which includes dialogue and learning (Knight, 2025). If GAI is to meaningfully contribute to academic research, domain-specific models aligned with rigorous scholarly values may be required. The recently released Claude for Education is an example of GAI tailored for higher education, in collaboration with institutions such as Northeastern University; but how effectively such models uphold ethical research standards remains an open question.

GAI as a research assistant

Recent attempts, for example from Andrew Maynard (Professor of Advanced Technology Transitions, Arizona State University), to test the ability of deep research tools to produce original research outputs have shown mixed results. suggests a balanced approach, highlighting the need for careful collaboration between humans and GAI. Rather than viewing GAI as a replacement, the academic community could explore how these technologies can complement and enhance traditional scholarly practices, maintaining the core values of academic inquiry.

GAI presents both opportunities and challenges. A key question is how academic publishers will adapt to its use in academic work. More importantly, it may affect academic careers and shift the roles of PhD students and post-doctoral researchers. Such developments will require informed responses from supervisors, senior academics and editorial boards.

GAI (ChatGPT-4o and MS Co-pilot) were used to edit parts of this text.

References

Cotton, D. R. E., Cotton, P. A., & Shipway, J. R. (2023). Chatting and cheating: Ensuring academic integrity in the era of ChatGPT. Innovations in Education and Teaching International, 61(2), 228–239. https://doi.org/10.1080/14703297.2023.2190148

Knight, S. (2025, February 17). From compliance to conversation: New guidelines push for ethical reflection in research reporting. BERA Blog. https://www.bera.ac.uk/blog/from-compliance-to-conversation-new-guidelines-push-for-ethical-reflection-in-research-reporting

Maynard, A. (2025). Does OpenAI’s Deep Research signal the end of human-only scholarship? Future of Being Human. https://futureofbeinghuman.com/p/openai-deep-research-ai-scholarship

Phan, L., Gatti, A., Han, Z., & Li, N. (2025). Humanity’s Last Exam. SuperIntelligence – Robotics – Safety & Alignment, 2(1). https://doi.org/10.70777/si.v2i1.13973  

Xia, Q., Weng, X., Ouyang, F., Lin, T. J., & Chiu, T. K. F. (2024). A scoping review on how generative artificial intelligence transforms assessment in higher education. International Journal of Educational Technology in Higher Education, 21(40). https://doi.org/10.1186/s41239-024-00468-z

Yan, L., Sha, L., Zhao, L., Li, Y., Martinez‐Maldonado, R., Chen, G., Li, X., Jin, Y. and Gašević, D. (2024). Practical and ethical challenges of large language models in education: A systematic scoping review. British Journal of Educational Technology, 55(1), 90–112. https://doi.org/10.1111/bjet.13370