Assessment of High Toxicity in Text Involving Women, Minorities, and Religious or Sexual Groups: A Case Study on Google's Perspective API

February 27, 2017

Google's Perspective API, a tool designed to evaluate the toxicity of online content, appears to grant elevated toxicity scores to text containing phrases related to women, minorities, religious groups, and sexual orientations. This observation raises concerns about AI bias and reinforces the importance of trustworthy and responsible AI governance. Ready to help us foster safe and secure AI? JOIN US This incident maps to the Govern function in HISPI Project Cerebellum Trusted AI Model (TAIM).
Alleged deployer
google
Alleged developer
google
Alleged harmed parties
women, minority-groups

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/13

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.