Feature Request: AI-Generated Content Transparency for Academic Integrity
Dear JASP Development Team, Dear community of JASP users,
First of all, thank you for the incredible work on JASP. It has become an essential tool in my statistics courses and for research in general. The recent integration of AI features is impressive and offers great potential for learning and exploration.
However, as a professor at the university level, I am facing a significant challenge regarding academic integrity and transparency. Many universities now have strict policies requiring students to disclose the use of AI in their assignments. Currently, when JASP’s AI generates annotations or interpretations, they appear identical to user-written text within the .jasp files.
There is no technical way for instructors to verify if a student has used the integrated AI, nor is there a mechanism for students to easily comply with institutional transparency requirements. Only a small "AI" text is reported in the title of the section (that is a good starting point), see image below.
I would like to suggest considering a feature that would ensure transparency for content generated by the integrated AI. Some possibilities could include:
- Automatic Disclosure Tags: A non-removable or default footnote/tag attached to any annotation or result section generated by the AI (e.g., "Interpretation generated by JASP AI on [Date].").
- Metadata/Log Transparency: Including a small, auditable metadata section in the
.jaspfile that lists the interactions with the integrated AI. - Encouraging "Best Practices": Designing the interface to explicitly prompt the student and the researchers to acknowledge the AI's contribution when generating text.
The goal here is not to "police" or restrict the use of AI, but to promote a culture of academic honesty and transparency, which is fundamental to scientific practice. Providing this feature would greatly help instructors support their students in using these powerful tools ethically.
I would be interested to hear what the community and the development team think about this. Is this something that could be considered for a future version of JASP?
Thank you for your time and for your continuous support of open science and education.
Best regards,
Michael
Comments
Dear Michael,
Thanks for the kind words and the suggestions. We will take these under consideration. One of the problems is that nobody knowns how this situation will evolve. As these tools become commonplace, it is not unthinkable that one or two years from now, AI will be treated just as any other tool, meaning that the researcher takes responsibility for the endproduct, but how the product came to be is no longer of interest (since everybody will effectively be using it).
As far as metadata are concerned, I am not a fan of including all interaction with the AI, if only for the reason that everybody should be allowed to hide their own ignorance. A student should feel uninhibited to ask "what does homoscedasticity mean?" without a permanent record of this being stored somewhere.
I could see the AI annotations use a different font, for instance. Then the student who wants to present the AI thoughts as their own would really have to go through some trouble. And perhaps a message somewhere (in the output?) that it is customary and recommended to report that you used AI?
EJ