On the way to responsible practices – Scope of AI
guidelines for scientific publishers and research data

Angelina Sophie Dähms, Linda Nierling, Felix Bach,
Christian Bonatto Minella, Florian Idelberger, Lea Sophie Singson

Most referenced publishers and guidelines with generative AI guidelines
(Source: Custom figure based on generative AI guidelines)
Generative artificial intelligence (AI) plays an increasing
role in science, since its 2022 launch. Large Language
Models (LLMs) are increasingly used, e.g. in literature synthesis,
hypothesis generation, code production, peer review support,
and research data annotation. This development does not merely
accelerate existing workflows. It alters epistemic conditions
under which scientific knowledge is produced, validated, and
disseminated.
Within the Leibniz ScienceCampus Digital
Transformation of Research (DiTraRe), these transformations
are analysed not only as technological innovation, but also with
regard to their impact on knowledge generation, workflows, outputs
and infrastructures.
Especially generative AI is seen to have the
potential to modify not only scientific knowledge, but to change scientific practice. Unlike previous computational tools, generative systems produce textual, visual, and analytical artefacts that resemble human-authored content.
This technical ability questions established criteria of authorship, reliability, accountability, and reproducibility.