Convenience vs Competence: The Evolving Role of Foundational Knowledge in AI-Driven Research

Vortrag mit anschließendem Sektempfang

Prof. Dr. Harald Sack (FIZ Karlsruhe – Leibniz-Institute for Information Infrastructure):
Convenience vs Competence: The Evolving Role of Foundational Knowledge in AI-Driven Research

The arrival of powerful Large Language Models (LLMs) — manifested by the quality leap seen with ChatGPT in late 2022 — has started a methodological revolution across science and research. The ability of these models to comprehend complex context has instantly disrupted established, multi-step toolchains in Natural Language Processing (NLP) and information extraction. These processes are increasingly being substituted by streamlined, end-to-end approaches that trans- form raw data directly into structured semantic knowledge. This fundamental shift offers substantial user benefits, enabling natural language interaction with research artifacts and thus bypassing the need for complex query languages. However, this convenience introduces a significant cost: a deep crisis of trust concerning the plausibility and veracity of LLM-generated results. More critically, we face an existential dilemma affecting both education and future research: How do we motivate the next generation of researchers to master foundational technologies — including programming, querying, data design, and ontological engineering — when LLMs seemingly negate this necessity? In this talk we will investigate this methodological change and its cascading effects on academic in- tegrity and fundamental skill development, drawing on concrete examples from our research spanning digital humanities (NLP and information extraction from historical documents), cross-disciplinary knowledge graph-based research data management, and the digitalization of materials science engineering.

Alle Interessierten sind herzlich eingeladen.

Zum Flyerpdf, 602 kb