The AMI Meeting Corpus is a multi-modal dataset comprising 100 hours of meeting recordings, being developed for meeting browsing technology. The corpus includes both naturally occurring and elicited meetings, with a focus on group dynamics and communication. It is recorded using various devices, including microphones, video cameras, and interactive tools, and is synchronized with each other. The data is hand-annotated for orthographic transcription, discourse properties, summaries, emotions, and gestures. The project aims to improve work group effectiveness and is expected to be valuable for research in speech, language, gesture, information retrieval, and organizational psychology. The corpus is structured to address methodological challenges by combining naturally occurring and elicited data, with the latter providing controlled conditions for experimental analysis.The AMI Meeting Corpus is a multi-modal dataset comprising 100 hours of meeting recordings, being developed for meeting browsing technology. The corpus includes both naturally occurring and elicited meetings, with a focus on group dynamics and communication. It is recorded using various devices, including microphones, video cameras, and interactive tools, and is synchronized with each other. The data is hand-annotated for orthographic transcription, discourse properties, summaries, emotions, and gestures. The project aims to improve work group effectiveness and is expected to be valuable for research in speech, language, gesture, information retrieval, and organizational psychology. The corpus is structured to address methodological challenges by combining naturally occurring and elicited data, with the latter providing controlled conditions for experimental analysis.