[TYPES] AI-generated conference submissions

[Mae Milano]

Hey Jon! Definitely interested in clearing up where the daylight is
here---and I think I have found it!  We might have different
estimations on what level of "understanding" students believe they
need before submitting work. The belief you are espousing here is
indeed the standard of our field---and the standard of our US/UK/EU
educational system more broadly.  But it is not the *universal*
standard with which all university students are acquainted! This
really surprised me too, the first time I ran into it.

There's two scenarios here---both sadly quite realistic.  The first
scenario is a student that has, through the examples of their faculty
mentors or their peers, been raised in a system that does not have a
rigorous attention to whether the student can defend their formalism.
To make an analogy to software engineering, this is the student that
just kinda throws more code at the problem until the compiler stops
complaining and most of the tests pass.  Can they explain *why* they
needed a type coercion here or there? Usually not! But it seems to do
what they expected, and they were always more interested in the "what"
than the "why."

The second scenario is a student who has *never* quite understood
formalism well enough but *believes that they do*. These students have
never really been held to a rigorous standard for formal development
at all.  To that student, they *already believe* they understand what
they are producing well enough to defend it---but are wildly
unprepared for the types of questions we would ask of such a
formalism, or the role that it is ultimately supposed to serve.  These
are the students I am *most interested* in reaching. These are also
the students that I have come across in my work and life, usually in
the form of international masters students who are, for the first
time, entering the "western" academic tradition.

I am not trying to defend the standards of the programs from which
they originated. In fact, I want to hold students, and all submitters,
to *our* standard.  But I also want to allow space for them to have
*not grown up in it*; I want to allow space for the things we will ask
them to do to amount to a real change in the way they are used to
thinking about their work more generally.

On Tue, Mar 17, 2026 at 12:16 PM Jon Sterling <jon at jonmsterling.com> wrote:
>
>
>
> On Tue, Mar 17, 2026, at 4:05 PM, Mae Milano wrote:
>
> Hey Jon,
>
> I think it's dangerous to conflate "uses of AI that were ultimately
> senseless" with "fraudulent generation of lies."  I was not aiming at
> students that *sought to deceive*: I was aiming at students who
> believed that their rule nests *supported their work*, and ultimately
> lacked the necessary training or insights to confirm this before
> submission. Such students may be "trying their luck," but they
> invariably believe that there's better-than-even odds of the stuff
> they've generated not just passing peer review, but actually being
> *right*.
>
>
> I wonder if we are speaking past each other... I think it is indeed a severe breach of ethics to send something that you knowingly haven't understood to a conference, and it is hard for me to come up with any valid argument to the contrary.
>
> I would also like to clarify that what matters is not whether the AI-generated stuff was senseless or sensible (indeed, that's almost entirely irrelevant)—what matters is that the author did not understand the output, but included it because they felt that is part of the convention of the field. That is a serious breach, and if I were aware of any students under my influence engaged in such things, they would receive a very firm reprimand. It would be a serious breach even if the AI-generated stuff turned out to be totally sensible. We should consider that possibility too (the senseless transmission of sensible things) when evaluating the ethics of these scenarios. It is indeed deeply wrong to submit text that you have not understood, whether or not that text contains valid or true science.