heiBOOKS

Abstract: The digital transformation of science is reshaping not only research workflows but also the integrity, openness, and societal trust in scientific knowledge. This paper investigates these developments through the lens of the Leibniz ScienceCampus DiTraRe and its interdisciplinary work on digital research infrastructures. Tracing the historical foundations of digital science – from Leibniz’s binary logic to the AI-driven research environments of today
– we highlight how shifts in data collection, knowledge organisation, and publication cultures redefine what constitutes scientific evidence and reproducibility. We examine contemporary challenges and potentials through the use case of the Chemotion Electronic Lab Notebook, demonstrating how domain-specific digital tools can foster transparency, efficiency, and trustworthiness across research lifecycles. A particular focus is placed on the role of artificial intelligence (AI), including generative models and large language models (LLMs), which are increasingly integrated into scientific processes. We discuss emerging practices such as multi-agent LLM collaboration to mitigate hallucinations and the rise of autonomous AI research assistants like Sakana AI’s “AI Scientist”. At the same time, the paper addresses the ethical and epistemic challenges posed by algorithmic processes, the impact of digitalisation on public trust, and the institutional responses through guidelines from DFG, UNESCO, and the European Commission. By connecting technical developments with cultural and normative reflections, we argue that building trusted digital research workflows requires a careful balance between innovation and responsibility, supported by interdisciplinary collaboration and continuous governance.  

Keywords: Generative AI, Open Science, Research Data, Reliability, Research Integrity