How Generative AI impacts research and teaching

1 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 How Generative AI impacts research and teaching A Journey Through Artificial Intelligence Lecture Robert Haase These slides can be reused under the terms of the CC-BY 4.0 license unless mentioned otherwise. CENTER FOR SCALABLE DATA ANALYTICS AND ARTIFICIAL INTELLIGENCE Funded by: http://doi.org/10.5281/zenodo.17846360 2 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 ScaDS.AI •part of Germany’s AI strategy •one of 6 AI centers in Germany •Technical focus on Big Data, Artificial Intelligence, HPC, Mathematical Foundations •Applications in Life Science, Medicine, Earth Science,… CENTER FOR SCALABLE DATA ANALYTICS AND ARTIFICIAL INTELLIGENCE http://scads.ai/ Map image modified from: https://ki-zentren.net/ 3 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Artificial Intelligence (AI) Unsupervised Machine Learning (ML) •Dimensionality reduction •Clustering •Outlier detection •Hypothesis generation Supervised ML •AI/ML learns tasks that need to be done by humans otherwise •Models are trained using annotated data Generative AI (GenAI) •Generates data from a prompt •Large Language Models •Training models is too expensive for individual projects •Hyped since 2022, with yet unclear limitations. UMAP 2 UMAP 1 Explorative Analytic Generative •Noise •Tourist jumpingg on a senso •Erthquake approaching! Explorative Analytic 4 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Overview Teaching Research Generative Artificial Intelligence 1 2 3 5 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Chat-APPs und Large Language Models Commercial LLM service providers https://gemini.google.com/https://chatgpt.com/ https://claude.ai/ 6 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Chat-APPs und Large Language Models Data protection / privacy compliant academic LLM providers (in Germany) and local models https://helmholtz-blablador.fz-juelich.de/ https://chat-ai.academiccloud.de https://ollama.com/ 7 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Chat-APPs und Large Language Models Specialized systems for academic research, e.g. summarizing publications Source: Generated from Mader N et al (2025), licensed CC-BY 4.0 https://osf.io/mp6rb_v1 8 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Chat-APPs und Large Language Models Specialized systems for academic research, e.g. summarizing publications Source: Generated from Mader N et al (2025), licensed CC-BY 4.0 https://osf.io/mp6rb_v1 9 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Chat-APPs und Large Language Models Specialized systems for academic research, e.g. summarizing publications Source: Generated from Mader N et al (2025), licensed CC-BY 4.0 https://osf.io/mp6rb_v1 16 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Critical review … is our job as brave [human] scientists Are these the mentioned 115%? Cropped from Li et al 2023, licensed CC-BY 4.0 https://arxiv.org/abs/2307.11760 If so, this is an improvement from very very bad to very bad 17 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Good scientific practice We humans are responsible for reviewing AI-generated content and to declare what LLM / GenAI-system was used and for what. World LLM Author LLM Author “ChatGPT was used for improving language in the manuscript, the intellectual content originates from the human authors.” 18 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Can LLMs solve programming tasks? Source: Jimenez et al (2024), licensed CC-BY 4.0 https://arxiv.org/abs/2310.06770 Plot data source: https://www.swebench.com/ (2025-06-21) 19 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 AI-generated Code … appears online more and more Indicator: Share of AIgenerated code to PythonContributions on Github.com (Hint: Presumable not submitted by AI-agents, but by humans using ChatGPT) Source: Daniotti et al (2025), licensed CC-BY 4.0 https://arxiv.org/abs/2506.08945 20 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Large-Language Models for Data Analysis Analysing images using ChatGPT Prompt: “Please segment the nuclei in the blue channel of the image” 21 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Collaborative Learning with AI Assistants •We all learn in parallel how to use GenAI •Many of us hide how we use it from others •Some are to shy to share how they use GenAI LLM LLM 22 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Collaborative Learning with AI Assistants Objectives: •Transparent peer-review of AI-generated content •human-to-human knowledge exchange how to use AI assistants https://github.com/haesleinhuepf/git-bob/ https://www.nature.com/articles/s43588-025-00781-1 H Domain expert H Data analyst Github / Gitlab CI Hi! I have data and would like to measure x. I propose to try algorithm alg. git-bob comment on how to do this with Python AI git-bob LLM service provider 1 LLM service provider 2 Certainly! Try this: from algorithms import alg x = alg(data) boxplot(x) For result visualization let’s better use a violinplot. git-bob ask claude to try this! I created a notebook for using alg and drawing a violinplot: LLM service provider 3 LLM gpt-4o LLM claude LLM gemini That looks ok/correct! Nice! Thanks! 23 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Collaborative Learning with AI Assistants •By introducing GenAI into our discussions, we could learn from each other, while using it transparently. LLM https://www.microsoft.com/insidetrack/blog/enabling-real-time-multiemployee-collaboration-at-microsoft-with-copilot-pages/ 25 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Generating teaching materials Enabling ad-hoc, asynchronous knowledge transfer https://generated-books.github.io/ AI-based, automatic generation of training materials AI-based, automatic translation of training materials 26 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Translating teaching materials “Please translate notebook <xyz.ipynb> to <language>.” https://scads.github.io/llm-bia2025/02_bia_bob/translating_notebooks.html 33 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Teaching strategies in the age of AI…. … require new forms of exams Pure online-tests are compromised. 34 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Teaching strategies in the age of AI…. … require new forms of exams Pure online-tests are compromised. Graded reports are compromised. Document: http://doi.org/10.5281/zenodo.17846360 Detector: https://scispace.com/ai-detector 36 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Good scientific practice “In view of its considerable opportunities and development potential, the use of generative models in the context of research work should by no means be ruled out. However, certain binding framework conditions will be required in order to ensure good research practice and the quality of research results: […] ➢It is in keeping with the professional ethics of researchers that they themselves should commit to compliance with the basic principles of research integrity. The use of generative models cannot relieve researchers of this content-related and formal responsibility.” Deutsche Forschungsgemeinschaft. (2025). https://www.dfg.de/resource/blob/289676/230921statement-executive-committee-ki-ai.pdf 37 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Good scientific practice ➢When making their results publicly available, researchers should, in the spirit of research integrity, disclose whether or not they have used generative models, and if so, which ones, for what purpose and to what extent. ➢Only the natural persons responsible can appear as authors in research publications. They must ensure that the use of generative models does not infringe anyone else’s intellectual property and does not result in scientific misconduct, for example in the form of plagiarism. Deutsche Forschungsgemeinschaft. (2025). https://www.dfg.de/resource/blob/289676/230921statement-executive-committee-ki-ai.pdf 38 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Good scientific practice ➢In decision-making processes, the use of generative models in/for proposals submitted to the DFG is currently assessed to be neither positive nor negative. ➢The use of generative models in the preparation of reviews is inadmissible due to the confidentiality of the assessment process. Documents provided for review are confidential and in particular may not be used as input for generative models. Deutsche Forschungsgemeinschaft. (2025). https://www.dfg.de/resource/blob/289676/230921statement-executive-committee-ki-ai.pdf Tools like Google’s NotebookLM are only allowed with published materials under permissive licenses 39 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Watch more… https://www.youtube.com/watch?v=UKa_3n8dQ2k https://www.youtube.com/watch?v=AOxHVvmdosQ https://youtu.be/oeH57siRwcU?si=0h-dClV7_R8yW7Kb https://youtu.be/9dtVlVwk2eg?si=c0wjqPUDCugVOoPO https://www.youtube.com/channel/ UCv7EDPwovn6T1iaMbhlpFjg And event more more ScaDS.AI Living Lab lectures 40 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Take-home Messages •Generative AI influences research and teaching increasingly •Summarizing literature •Scientific literature research •Teaching material production (text, code, images, …) •The human authors are responsible for contents •Uploading private / secret documents to LLM services providers is prohibited •Important: Critical review of •Literature proposed by AI •AI-generated content •AI-related content (“AI Snake Oil”) 41 Impact of Generative AI Arqus Living Lab Lecture Robert Haase @haesleinhuepf December 8th 2025 Thank you for your attention! Contact Dr. Robert Haase ScaDS.AI Dresden/Leipzig Universität Leipzig Humboldtstraße 25 04105 Leipzig [email protected] These slides can be reused under the terms of the CC-BY 4.0 license unless mentioned otherwise. http://doi.org/10.5281/zenodo.17846360