Understanding the emergence of AI in the training and supervision of our doctoral students: what about academic integrity?

The objective of the study carried out at the Institute for Research and Action on Fraud and Academic Plagiarism (IRAFPA) in Geneva in Switzerland is to take stock of the perception that doctoral students have of generative AI, when they were unfamiliar with these tools during their master’s studies. The aim is also to clarify the use of these tools in their daily work and the attitude of tutors in this regard.

The aim is to offer these two audiences – doctoral candidates and supervisors – a reliable, transparent, generous, useful speech… Therefore, a formation of integrity in continuous evolution. But the extremely rapid change in AI structurally clashes with the length of training of future researchers: three years of doctorate is short to root academic integrity in everyday practices.

Certainly, the solution is not a mandatory MOOC for doctoral students (which has the merit of existing) or the fact of delegating education to integrity to ethics specialists «outside the earth». And the thesis supervisors are not necessarily the best placed, while the directors of doctoral schools observe that the diploma of Authorization to Conduct Research (HDR2) is given to people who have not received any training in integrity.

Our individual and collective responsibility is anchored in this question of knowing how to train their doctoral students in doubt, in a critical spirit, in the spirit of synthesis that will be needed more than ever.

An open survey was conducted among doctoral candidates and doctoral thesis directors from French universities. We analyzed 115 responses from doctoral candidates and 40 from thesis supervisors. The responses of a complementary sample of thirty-three directors of doctoral schools were then analyzed to ascertain the means available to them to mobilize tutors on general integrity issues. They responded extensively to our open questions analyzed by inductive method of sense extraction.

Our survey reveals that many professors are between denial and fear, and that they delegate to other instances to take into account this new situation. The parallel of the leak to the software of detection of textual plagiarism is the very example of this fear of the mobilization of some and others.

So, in the face of the AI revolution – and although training in academic integrity remains superficial, our preliminary results show that we must return to the fundamentals of pedagogy:

  • It was found that doctoral students were seeking explicit, not tacit competences.
  • Supervisors are sensitive to a pedagogy of the experience (or multiple experiences) of members of the academic body (cf. The theatre of situations).
  • Heads of doctoral schools often lack the means and time to understand the ongoing phenomenon.

It remains to define the action of acculturation to integrity by assimilation in a world disturbed by AI. Indeed, IRAFPA is working tirelessly to ensure that this acculturation does not proceed by “segregation” (ethics courses in the field taught by “specialists”), by “risky integration” (at the will of the different managers), and much less by “marginalization” (there is no talk of integrity and ethics). But how to define and implement this “acculturation to integrity by assimilation”?

On this subject, a colloquium on Academic Integrity «Challenges and uncertainties of academic integrity in the era of artificial intelligence» is organized at the University of Coimbra (Portugal), on 20 – 22 June 2024. See https://irafpa.org/colloqui- 2024coimbra/presentation-2024/ . USF will participate.
Contact: michelle.bergadaa@unige.ch.

Robert Laurini

Editor Professor Emeritus in Information Technologies