Tagged with: gamelan

New AI tools offer potential benefits for transcribing, classifying, and summarizing large collections of fieldwork notes and recordings. Services like mygoodtape.com and Deepgram’s AI speech-to-text API can provide low-cost transcription. The raw output can then be cleaned up and summarized using large language models like Anthropic’s Claude.

By explaining field-specific jargon to the AI system, the accuracy of transcription and summarization can be further improved. With just these basic tools, it may be possible to get 90% of the way towards generating publishable content from large volumes of fieldwork data. The time savings in transcription and synthesis could be considerable.

Overall, AI looks promising for streamlining analysis and publication based on qualitative fieldwork, but care and human judgment is still needed to ensure accuracy and interpretability.