The main objective of TrustLLM is the development of an open, trustworthy, and sustainable LLM initially targeting the Germanic languages. This will create the foundation for an advanced open ecosystem for next generation modular and extensible European trustworthy, sustainable, and democratized large language models. A lot of focus is put on trustworthiness, aiming to enhance and refine factual correctness, thereby addressing current challenges with misinformation and hallucinations. At a national level, GPT-NL is a collaborative venture by Dutch organizations TNO, SURF, and NFI, focusing on creating a Dutch large language model that aligns with European values such as transparency, data privacy, and intellectual property rights. GPT-NL will be constructed from scratch, prioritizing GDPR-compliant data use and stringent intellectual property rights management to prevent copyright infringement. This initiative not only showcases the responsible creation of sovereign models in alignment with legislation but also integrates principles of responsible AI directly into the development process, aiming to become a benchmark for developing future generative AI systems that are both democratic and trustworthy.
Speaker: Saskia Lensink (TNO)
