Researchers turn water into a living AI sensor

Researchers Kuan-Ju Wu, Youyang Hu, Chiaochi Chou, and Yasuaki Kakehi have developed **Sensus Pond**, a radical human-computer interaction system that reimagines water as an active sensing medium for more-than-human observation. Published in *Proceedings of IASDR2025*, this method challenges anthropocentric design by avoiding translation or classification of nonhuman traces. Instead, it uses **Swept Frequency Capacitive Sensing** combined with an **Artificial Neural Network** to register subtle, ephemeral interactions at the water’s surface—such as ripples from insects, amphibians, or wind.

The system doesn’t decode or label these events. Instead, it visualizes temporal accumulations of overlapping traces, creating a layered archive of ecological activity. By emphasizing ‘attunement over control,’ Sensus Pond shifts the designer’s role from interpreter to facilitator, enabling open-ended multispecies encounters. The result is a poetic reflection on shared ecological life—where ambiguity and contingency become part of the design aesthetic.