Programming, outreach, and AI use

Programming

I have written several small programs to guide or visualize my research, for example:

  • Generator for cards for Dobble (AI supported): Generate your own dobble cards
  • GAP program to generate the Veech group of a cover of the Chamanara surface (with Gabi Weitze-Schmithüsen): Chamanara covers
  • Tool in SageMath that visualizes the translation structure that is given by a differential on a subset of the complex plane

Outreach

I regularly give talks and workshops for a broad audience. Examples include:

  • Workshop and talk at the Museum for minerals and mathematics on the mathematics behind the game Dobble
  • Talk at the Akademische Mittagspause: youtube video (in German)
  • Introducing Maryam Mirzakhani's work at a movie evening in Karlsruhe
  • Talks at the Day of Mathematics for high school students in Karlsruhe and in Saarbrücken
  • Invited expert for an after-movie discussion for "Flatland" at a film festival

I am also the senior editor of the outreach project Snapshots of modern mathematics from Oberwolfach.

Geometry lab

I was involved in the founding of the Heidelberg Experimental Geometry Lab (HEGL), the first European member of Geometry Labs United.

AI statement

I think that the costs of generative AI (regarding environment, society, democracy, ...) outweigh the benefits. However, the situation and my opinion is very much evolving, so this statement will always reflect the status at one point in time (currently October 2026).

Previous use

Previously, I have tried out free AI models and have sometimes been carried away by excitement and sometimes been severely disappointed. Programming the generator for Dobble cards above has been an exciting experience - and the only one that was published in any way.

I have not discussed any of my research outputs with generative AI and none of the texts I have published has been edited by AI on my side. I have not discussed this question with all of my collaborators, so I cannot claim that no generative AI has been used at all.

Future (non) use

I have chosen to not use generative AI as a source of ideas in my research, even if it slows down my research (seen relatively, compared to the increase in speed for AI generated results). I also try to divert my attention from AI generated "research" to human mathematics as much as possible.

I do, however, use AI right now (October 2026) for automatic translation as I recently moved to a country whose language I am not fluent in yet and where I have to handle a lot of bureaucracy. I will continue doing so for the next months until a simple dictionary will be enough.

Should I ever decide to use AI support for any stage of a research article, even for proof-reading, I will declare this. From now on (October 2026), I will also ask my collaborators about their AI use, so it can be made transparent.