Sound Intelligence's Aggression Detector Misidentified Innocuous Sounds

June 25, 2019

Sound Intelligence's AI-based aggression detection algorithm, deployed in schools, reportedly exhibited a significant number of false positives, misclassifying everyday sounds such as laughter, coughing, cheering, and loud discussions.

Contributing to safe and secure AI practices, join us in shaping the future with Project Cerebellum by addressing concerns like this through our Harm Incident Mapping (HISPI Project Cerebellum TAIM).
JOIN US

Matched TAIM controls

Suggested mapping from embedding similarity (not a formal assessment). Browse all TAIM controls

MEASURE 2.6 — similarity 0.674, rank 1. TAIM detail and related incidents →
MAP 5.1 — similarity 0.651, rank 2. TAIM detail and related incidents →
MAP 4.1 — similarity 0.648, rank 3. TAIM detail and related incidents →

Alleged deployer: rock-hill-schools, pinecrest-academy-horizon
Alleged developer: sound-intelligence
Alleged harmed parties: students, rock-hill-school-students, pinecrest-academy-horizon-students

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/404

Data source

Incident data is from the AI Incident Database (AIID).

When citing the database as a whole, please use:

McGregor, S. (2021) Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database. In Proceedings of the Thirty-Third Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-21). Virtual Conference.

Pre-print on arXiv · Database snapshots & citation guide

We use weekly snapshots of the AIID for stable reference. For the official suggested citation of a specific incident, use the “Cite this incident” link on each incident page.