At the 11th International Summer School on AI and Big Data, Dr. Nicolas Flores-Herr (Fraunhofer IAIS) will give a keynote speech on Multilingual Large Language Models: How Can Science boost European Sovereignity, Open Source and the Economy?

Keynote: Multilingual Large Language Models: How Can Science boost European Sovereignity, Open Source and the Economy?

Europe stands at a pivotal moment in AI development. As the OpenGPT‑X consortium recently demonstrated, Teuken‑7B, a 7 billion‑parameter model trained from scratch on all 24 official EU languages with over 50 % non‑English data, is a landmark achievement for European digital sovereignty. Available under the Apache 2.0 license it offers users transparent, locally controlled AI that respects data privacy and fosters homegrown innovation.

In this keynote, Dr. Nicolas Flores-Herr argues that scientific leadership in multilingual LLMs is critical on three fronts: sovereignty, open source, and economic impact. First, science-driven, multilingually pre‑trained models like Teuken not only preserve Europe’s communications diversity but also fuel adaptability and trust in AI systems. Second, by embracing established OSI‑approved licenses, Europe can accelerate adoption: legal clarity matters, especially learned from OpenGPT‑X’s commercial integration via Apache 2.0. Third, investing in open‑source multilingual LLMs catalyzes a robust AI ecosystem, enabling European startups, SMEs, and public institutions to fine‑tune, deploy, and innovate on sovereign AI infrastructure.

To amplify these gains, Dr. Nicolas Flores-Herr will spotlight the work on Teuken (a project on instruction‑tuned European multilingual LLMs) as a future track in the OpenGPT‑X roadmap, reinforcing how scientific rigor and open‑source best practices can multiply economic spillovers across industries. Together, they form the foundation for a genuinely European AI ecosystem: competitive, sovereign, inclusive, and open.

Dr. Nicolas Flores-Herr

Dr. Nicolas Flores-Herr is Team Lead Foundation Models & Gen AI Systems at the Dresden site of Fraunhofer IAIS. He leads a team that develops customized generative AI solutions for a wide range of industries – from enterprise chatbots to knowledge discovery tools. The team uses both established frameworks and its own RAG technology stack to realize innovative AI solutions.

He is also involved in the development of large language models and multimodal AI models: from training completely new models to further pre-training, fine-tuning and instruction tuning. Much of this work takes place within the OpenGPT-X, TrustLLM, and Eurolingua projects.

Dr. Nicolas Flores-Herr’s academic career included a PhD in physics, as well as ten years of research experience in neuroscience and cell physiology. Today, he uses this scientific perspective to develop AI solutions that help companies innovate and grow.