Inappropriate AI Response in Google's Gmail Smart Reply Tool: A Case Study on Safe and Secure AI

November 3, 2015

Google's Gmail Smart Reply tool, designed to enhance user experience, was found suggesting inappropriate responses such as 'I love you'. This incident underscores the need for trustworthy AI and highlights the importance of responsible AI governance in preventing harm. Join us in shaping safer AI landscapes with Project Cerebellum, where we strive to establish guardrails for AI like this AI incident which maps to the Govern function in HISPI Project Cerebellum Trusted AI Model (TAIM). JOIN US
Alleged deployer
google
Alleged developer
google
Alleged harmed parties
gmail-users

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/17

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.