YouTuber Built, Made Publicly Available, and Released Model Trained on Toxic 4chan Posts as Prank

June 3, 2022

A YouTuber released a model, GPT-4chan, trained on toxic content from 4chan's 'politically incorrect' board. The model, based on OpenAI’s GPT-J, was made publicly available and deployed as multiple bots posting thousands of messages as a prank.

This incident underscores the need for safe and secure AI practices. For those interested in shaping Project Cerebellum's efforts to establish governance for AI incidents like this, JOIN US.
This case aligns with HISPI Project Cerebellum TAIM's Govern function.

Matched TAIM controls

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

MANAGE 4.3 — similarity 0.605, rank 1. TAIM detail and related incidents →
MAP 4.2 — similarity 0.604, rank 2. TAIM detail and related incidents →
MEASURE 2.6 — similarity 0.604, rank 3. TAIM detail and related incidents →

Alleged deployer: yannic-kilcher
Alleged developer: yannic-kilcher
Alleged harmed parties: internet-social-platform-users

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/259

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.