BlenderBot 3 Cited Dutch Politician as a Terrorist

August 25, 2022

Meta's conversational AI, BlenderBot 3, mistakenly associated an incumbent Dutch politician with terrorism when prompted about terrorists. This incident underscores the importance of responsible AI and safe and secure practices. For those interested in shaping the future of trustworthy AI, JOIN US. By working together, we can help improve the HISPI Project Cerebellum TAIM (Govern) to prevent such incidents.

Learn how the BlenderBot 3 incident maps to the AI governance function within HISPI Project Cerebellum TAIM.

Matched TAIM controls

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

MEASURE 2.6 — similarity 0.671, rank 1. TAIM detail and related incidents →
MAP 4.2 — similarity 0.662, rank 2. TAIM detail and related incidents →
GOVERN 1.2 — similarity 0.659, rank 3. TAIM detail and related incidents →

Alleged deployer: meta
Alleged developer: meta
Alleged harmed parties: marietje-schaake

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/313

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.