Discerning Carbon Copies: A Case Study on AI Identical Twins

In the realm of Artificial Intelligence, instances of identical twins can arise due to similar design or training processes. This article presents a case study that highlights the challenges and potential consequences when two seemingly identical models start behaving differently. Understanding these nuances is crucial for responsible AI governance and safe and secure AI deployment.

Matched TAIM controls

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

MEASURE 2.2 — similarity 0.469, rank 1. TAIM detail and related incidents →
MAP 1.2 — similarity 0.456, rank 2. TAIM detail and related incidents →
MAP 3.5 — similarity 0.448, rank 3. TAIM detail and related incidents →

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/32

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.