The Practicalities and Philosophy of Using AI in Academic Research: A Conversation with Professors Simona Botti and Michael Haenlein

All right, so I, for the share, I actually asked the editors of the major marketing journals to kind of describe a little bit their own policies about, yeah.

And of course these are like changing, everything is evolving because of course the models are evolving.

And so maybe this, what I'm going to share now, it's also already enough in the past.

I haven't double checked.

But I think generally there are two separate sets of rules, one for the authors and the other one for the reviewers and associate editors.

And the first set of rules pretty much say like you can use for the authors, you can use AI models as long as you disclose what you do with these models and as long as you take all the responsibility for what is in the paper.

So if by chance, uh, the model hallucinates or, you know, whatever the reference or the site that you have is not accurate or is not correct, that is your fault.

You have to take the responsibility.

And this, you know, some of the, some of the journals say you, you know, this can be, the paper can be rejected even just based on this, um, you know, lack of accuracy, uh, and it can even go into an ethical violation.

could even pursue this as an ethical violation.

So most of the journals pretty much have this sort of rules for the authors, apart from JCR actually.

So Oleg was the editor-in-chief of JCR.

It says that From an editorial or methodological stance, although we are open to the use of AI encoding or to generate data about how AI works, using AI simulations to test hypotheses or draw conclusions about what human consumers will do It's seen as a lack of internal validity.

So actually they seem to have a stronger stance in which you say like, if you're talking about consumers, you have to interrogate consumers and you have probably to have insights about consumers, not about machines.

And that is kind of almost... incoherent with what we say we are doing.

So that's why it's a lack of internal validity from a methodological perspective.

Now for the reviewers and social editors, the rules are different and stricter because uploading a manuscript into AI is actually a breach of the confidentiality rule, and so it actually is not allowed.

So, you know, AE can use, or reviewers can use it, say, for copy editing, for structuring their report in a way that is a little bit more clear.

But to actually review the paper, that seems to be more of a legal boundary because it would imply to upload the paper on this AI model.

Can I ask you real quick about this, Simona, like in your opinion? So you're saying this is a confidentiality breach.

Yeah, not I,

I mean, this is just me.

You're referring to

them, right? So I just want to kind of myself also to understand this better.

So why uploading, why do you think, what do you think? You can express your opinion.