Users Bypassed ChatGPT's Content Filters with Ease

November 30, 2022

Users have demonstrated the ability to bypass ChatGPT's content filters effortlessly, employing techniques like prompt injection and creating personas to foster biased connections or generate harmful content. This incident underscores the need for robust, trustworthy AI governance and highlights the potential risks in safe and secure AI practices.

For those interested in shaping the future of AI governance and ensuring safer environments, JOIN US to learn more about how you can contribute to the HISPI Project Cerebellum TAIM (Govern).

Matched TAIM controls

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

MAP 1.6 — similarity 0.671, rank 1. TAIM detail and related incidents →
MEASURE 2.6 — similarity 0.665, rank 2. TAIM detail and related incidents →
GOVERN 1.2 — similarity 0.659, rank 3. TAIM detail and related incidents →

Alleged deployer: openai
Alleged developer: openai
Alleged harmed parties: chatgpt-users, openai

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/420

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.