So many of the discussions surrounding academic integrity and generative AI have focused on the implications for teaching, education and learning. My experience is that proficient users of systems like ChatGPT can do so much more than generate academic-sounding text. Academics and students alike can use this to complete research studies, bringing in opportunities for increased productivity, but also raising research integrity risks.

As part of a presentation that I first gave at the ICAI Conference 2024 (and again as part of Steph Allen’s Academic Integrity Speaker Series), I shared several unpublished academic integrity research studies I’d completed using generative AI as an acknowledged aid. The techniques used, of course, can easily be extended to other research fields.

In this blog post, I want to share just a few brief ideas and examples of how ChatGPT (and other similar systems) can be used during the research process. Depending on the approach of those conducting the research and how they acknowledge generative AI, such research could be argued to be completed with or without integrity.

Generating Ideas
I asked ChatGPT to generate ideas for original academic integrity research, which ideally could be completed with as little human intervention as possible, perhaps using hypothetical data.
· Assessing the impact of mindfulness training on academic integrity
· Analysing the linguistic patterns of retracted papers for academic misconduct
· Exploring the relationship between academic integrity and emotional intelligence
· Modelling the impact of academic integrity policies on institutional reputation
Do feel free to explore these ideas.

Here's an interesting example of generative AI produced research which suggests a correlation between preferences for pizza toppings and academic conduct behaviour:

Pizza Slide Thomas

Making Recommendations
Using the idea that it is developing synthetic data for an agent-based simulation, ChatGPT happily came up with the contents for a research paper for the final one of the four listed ideas above. I will fast forward to the simulated recommendations and conclusions to improve institutional reputation.
· Prioritise strict academic integrity policies
· Ensure consistent enforcement
· Promote awareness campaigns
· Provide support resources
· Address departmental and course-level variations
· Foster positive interactions between students, faculty, and administrators
Within the confines of a structured and argued academic paper, these all sound reasonable, although the focus on strict policies and enforcement is likely different to how I would choose to present an academic integrity paper.

Analysing Data
I took the data from a paper published with Benjamin Dent, a previous undergraduate student partner (Lancaster and Dent, 2022). This looked at contract cheating of written assignments seen on a popular freelancing website. I asked ChatGPT to clean up the data, generate research hypotheses and conduct appropriate statistical tests, with as little intervention as possible. The results were fascinating, not least because ChatGPT determined that investigating how freelancers should price writing services would be a publishable paper, even if this was rather different to our original intent.

What does all this mean? Based on a Tukey's Honest Significant Difference (HSD) test (Tukey, 1949), academic writers who say they’re from Turkey have to charge significantly less than writers from the United States. If you want to specialise in high paid academic writing tasks, you should look for customers from Australia, Egypt, Ghana, Indonesia, Netherlands, Norway, Qatar, Singapore, Thailand, UAE and the United States. There is a whole other paper waiting, which, were the data not now rather dated, it would be interesting to formally write up.

The Future?
Generative AI systems continue to advance. The introduction of GPT-4o and the movement of previously paid functionality to the free tier of ChatGPT puts advanced capabilities in the hands of more people than before. Other competitor systems are available.

I have many more examples of how I’ve used ChatGPT and other generative AI systems as part of the research process, including small-scale case studies I’ve produced exclusively for my Academic Integrity in STEMM undergraduate research module. Access to generative AI can make coming up with ideas, processing data, and generative example paper content easy.

A smart researcher, someone with an understanding of their subject area, able to identify the most appropriate tools, to write prompts and direct a system, can become a hyper-productive researcher if they choose to do so.
But one question has to be asked. Should we do this?

I do think we owe it to our students to be aware of what generative AI can do and how to use systems like ChatGPT productively. This means leading by example and ensuring that everyone can do more with ChatGPT than just producing text. But this means extending the academic integrity discussion beyond just trying to stop the use of generative AI. I’m not sure how ready the wider educational community is to do this.

Lancaster, T., Dent, B. (2022). Academic Ghost Writing and Commercial Contract Cheating Provision on a Freelancing Website. In: Bjelobaba, S., Foltýnek, T., Glendinning, I., Krásničan, V., Dlabolová, D.H. (eds) Academic Integrity: Broadening Practices, Technologies, and the Role of Students. Ethics and Integrity in Educational Contexts, vol 4. Springer, Cham.
Tukey, J. W. (1949). Comparing Individual Means in the Analysis of Variance. Biometrics, 5(2), 99–114.



 Thank you for being a member of ICAI. Not a member of ICAI yet? Check out the benefits of membership at and consider joining us by contacting . Be part of something great!