Personal voice assistants struggle with black voices, new study shows

March 23, 2020

A recent study exposed a concerning bias in voice recognition systems from major tech companies, including Apple, Amazon, Google, IBM, and Microsoft, which consistently misinterpreted speech from black speakers. This underscores the urgent need for safe and secure AI practices.

Join us at Project Cerebellum to help govern, map, and measure harm prevention strategies in voice recognition systems through our Harm Incident Mapping Initiative (HISPI Project Cerebellum TAIM). Together, we can build guardrails for AI and shape a future of responsible AI.

JOIN US

Matched TAIM controls

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

MAP 1.6 — similarity 0.669, rank 1. TAIM detail and related incidents →
MEASURE 2.10 — similarity 0.658, rank 2. TAIM detail and related incidents →
MAP 4.1 — similarity 0.655, rank 3. TAIM detail and related incidents →

Alleged deployer: microsoft, ibm, google, apple, amazon
Alleged developer: microsoft, ibm, google, apple, amazon
Alleged harmed parties: black-people

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/102

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.