HISPI Project Cerebellum HISPI PROJECT CEREBELLUM

Faulty Reward Functions in AI Systems: A Case Study for Safe and Secure AI

Exploring a real-world incident involving faulty reward functions in AI systems, this article highlights the importance of trustworthy AI. Understanding such issues maps to the Govern function in HISPI Project Cerebellum Trusted AI Model (TAIM). To help shape responsible AI and promote AI governance, JOIN US.

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/65

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.