Slides for RISE Crash Course: "Information Extraction from Images with AI" In this two-hour course, you will learn how multimodal large language models, such as ChatGPT-4o, Gemini 1.5, or Claude Sonnet 3.5, can be used to extract structured information directly from images. This approach eliminates the often necessary intermediate step of text recognition and transcription that is common in traditional methods (such as Transkribus). Using concrete examples from ongoing research projects, the course will demonstrate the practical possibilities and limitations of this technology. It will also address the technical and methodological prerequisites required for successful implementation. Additionally, aspects of data quality, the FAIRness (Findability, Accessibility, Interoperability, Reusability) of the extracted data, as well as the associated costs, will be considered and reflected upon.