Chatbot in Workplace Training at Bunbury Prison Reveals Real Names in Sexual Harassment Case

August 20, 2024

During training sessions at Bunbury Prison in Western Australia, Microsoft's Copilot AI chatbot was employed to generate case study scenarios. Regrettably, the chatbot disclosed confidential information by incorporating the real name of a former employee implicated in a sexual harassment case. This underscores the importance of trustworthy AI and robust AI governance.

For those interested in shaping the future of safe and secure AI practices, join us at HISPI Project Cerebellum TAIM to Govern, Map, Measure, or Manage such incidents, ensuring AI adheres to established guardrails for harm prevention. JOIN US

Matched TAIM controls

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

GOVERN 2.2 — similarity 0.623, rank 1. TAIM detail and related incidents →
MAP 1.6 — similarity 0.622, rank 2. TAIM detail and related incidents →
MEASURE 2.10 — similarity 0.618, rank 3. TAIM detail and related incidents →

Alleged deployer: charlotte-ingham, western-australia-department-of-justice
Alleged developer: microsoft
Alleged harmed parties: bronwyn-hendry, western-australia-department-of-justice, western-australia-department-of-justice-senior-staff-members

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/773

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.