Richard Schmidt's research utilizes the use of large language models (LLMs) as extractive and analytical tool, enabling qualitative research to be conducted at a quantitative scale. While these methodologies open new frontiers for analyzing complex social and scientific phenomena, his work critically deploys rigorous reliability assessments. This focus extends beyond technical accuracy and alignment, addressing broader social and business implications to ensure the robust and responsible integration of LLMs into society.

Richard's academic journey began with a Bachelor’s degree in Culture and Technology at TU Berlin, where he explored social, political, and economic disciplines alongside interdisciplinary research methods. His undergraduate thesis employed natural language processing (NLP) to conduct sentiment analysis of German newspaper coverage on China – a culmination of his China studies at the pronoun Technical China Center of the TU Berlin. Pursuing a Master’s in Data Science and Society at Tilburg University, Richard deepened his technical expertise in machine learning while maintaining a focus on socially relevant applications. His master’s thesis, centered on graph neural networks for spatio-temporal pollution prediction, exemplifies his commitment to leveraging advanced methodologies for addressing environmental and societal challenges. Both his undergraduate and graduate studies provided a foundation in NLP techniques, including sentiment analysis, named entity recognition, and automated speech recognition.

Richard’s growing interest in the rapidly evolving domain of LLMs has been shaped by a critical approach to their applications and limitations. His curiosity and commitment to trustworthy AI systems led him to concentrate on alignment and evaluation methodologies, culminating in his current role at the Quantitative Science and Technology Studies Team at VTT Technical Research Centre of Finland. There, he contributes to the development and reliability assessment of language models as tools for research.

Driven by a dedication to technological accountability, Richard engages with the regulatory and societal responses to the uncertainties surrounding LLMs, particularly their potential misuse in the absence of robust technical safeguards. His work seeks to enhance the transparency and explainability of these models, fostering a more reliable and equitable technological future.