Datascapes: Sonification for Anomaly Detection in AI-Supported Networked Environments ↗
paperFrontiers in Computer Science · January 2024
- Proposes a replicable design framework for sonifying anomaly detection in AI-monitored networks, tested on both a water distribution system and corporate internet infrastructure
- Evaluated by eight cybersecurity experts in real-world conditions, confirming that continuous soundscape metaphors let operators detect anomalies peripherally without adding to the already overloaded visual channel
- Argues explicitly that in real-time monitoring of networked systems, new solutions must not increase the burden on the visual channel — the ear is the underused observability interface