Towards AI-Supported Research: a Vision of the TIB AIssistant

Towards AI-Supported Research: a Vision of the TIB AIssistant Sören Auer1,2, Allard Oelen1, Mohamad Yaser Jaradeh2,1, Mutahira Khalid1, Farhana Keya1, Sasi Kiran Gaddipati1, Jennifer D’Souza1, Lorenz Schlüter3, Amirreza Alasti3, Gollam Rabby2, Azanzi Jiomekong4,1 and Oliver Karras1 1TIB – Leibniz Information Centre for Science and Technology, Hannover, Germany 2L3S Research Center, Leibniz University of Hannover, Hannover, Germany 3Leibniz University of Hannover, Hannover, Germany 4University of Yaounde 1, Yaounde, Cameroon Abstract The rapid advancements in Generative AI and Large Language Models promise to transform the way research is conducted, potentially offering unprecedented opportunities to augment scholarly workflows. However, effectively integrating AI into research remains a challenge due to varying domain requirements, limited AI literacy, the complexity of coordinating tools and agents, and the unclear accuracy of Generative AI in research. We present the vision of the TIB AIssistant, a domain-agnostic human-machine collaborative platform designed to support researchers across disciplines in scientific discovery, with AI assistants supporting tasks across the research life cycle. The platform offers modular components — including prompt and tool libraries, a shared data store, and a flexible orchestration framework — that collectively facilitate ideation, literature analysis, methodology development, data analysis, and scholarly writing. We describe the conceptual framework, system architecture, and implementation of an early prototype that demonstrates the feasibility and potential impact of our approach. Keywords AI-Supported Research, LLMs for Science, Scholarly AI Platform, Scholarly Research Assistant 1. Introduction The developments of Generative AI, and specifically Large Language Models (LLMs), have had a significant impact in many areas of our society [1] already. Additionally, in the scientific domain, LLMs are increasingly utilized, for example, to assist researchers with academic writing [2]. LLMs are used across a wide variety of scholarly domains, such as medicine in life sciences [3], social sciences [4], chemistry [5], law [6], and coding in computer science [7]. While many approaches have been proposed and demonstrated, it remains challenging for researchers to get started with LLMs in their field. The ability of individual researchers to optimally leverage AI for their research strongly depends on their AI literacy, i.e., their ability to evaluate, communicate with, and collaborate using AI [8]. AI can be used to support domain-independent tasks, such as finding related work, assisting with paper authoring, and proofreading, as well as for domain-specific tasks, including supporting methodologies, implementations, or evaluations. While the possibilities are virtually unlimited, the benefits that a single researcher gains from AI-assisted research heavily depend on the user and the specifics of her research work, which determine the required prompts and the tasks that can be outsourced to the LLM. Prompt engineering is a crucial aspect for effective LLM 5th International Workshop on Scientific Knowledge: Representation, Discovery, and Assessment, Nov 2024, Nara, Japan $ auer@tib.eu (S. Auer); allard.oelen@tib.eu (A. Oelen); jaradeh@l3s.de (M. Y. Jaradeh); mutahira.khalid@tib.eu (M. Khalid); farhana.keya@tib.eu (F. Keya); sasi.gaddipati@tib.eu (S. K. Gaddipati); jennifer.dsouza@tib.eu (J. D’Souza); lorenz.schlueter@stud.uni-hannover.de (L. Schlüter); amirreza.alasti@stud.uni-hannover.de (A. Alasti); gollam.rabby@l3s.de (G. Rabby); jiomekong@tib.eu (A. Jiomekong); oliver.karras@tib.eu (O. Karras)  0000-0002-0698-2864 (S. Auer); 0000-0001-9924-9153 (A. Oelen); 0000-0001-8777-2780 (M. Y. Jaradeh); 0000-0001-8482-4004 (M. Khalid); 0000-0000-0000-0000 (F. Keya); 0000-0003-3098-4592 (S. K. Gaddipati); 0000-0002-6616-9509 (J. D’Souza); 0009-0002-1165-773X (A. Alasti); 0000-0002-1212-0101 (G. Rabby); 0000-0002-8005-2067 (A. Jiomekong); 0000-0001-5336-6899 (O. Karras) © 2025 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). Prompt library Server 1 Server 2 ... Researchers Domain experts AI experts Developers Community driven TIB AIssistant platform Tool library Data store Core system modules MCP servers System actors System design principles External Human-machine collaboration Personalization Customizability Transparency Error-tolerance Flexibility Figure 1: Proposed framework of AI-assisted research, highlighting system actors, external MCP (Model-Context Protocol) servers consisting of collections of tools, core system modules, and the system’s design principles. usage [9], but can be a bottleneck for non-AI experts [10]. Even if researchers possess the necessary skills to operate LLMs, research work relies on diverse tools designed to support specific tasks, such as data analysis in computing environmen

This content is AI-processed based on open access ArXiv data.