01 02 03


In response to the concept of a "non-human theatre," we created a system that uses an artificial intelligence model to analyze rehearsal scenes, notes, and photos of interviews and display the results. The model can identify about 1000 objects, but there is a bias in what is being learned and the results are also returned in English. It is converted to Japanese by the English-Japanese dictionary, so the answer may sound strange to a Japanese speaker, but it works correctly as a program. The aim was to highlight the differences and misalignments in perceptions with humans by allowing a "biased artificial intelligence" to appreciate it, rather than seeking accuracy in analysis.