← the record
AIAAIC-0266

GPT-3 bot posts Reddit comments unnoticed

A bot powered by OpenAI's GPT-3 large language model spent a week responding to comments on the Ask/Reddit subreddit before it was discovered not to be human. Though the bot posted mostly harmless feedback, it also engaged with conspiracy theories and sensitive topics, including suicide. In addition to angering the Reddit commmunity, the incident was seen to show the ease with which GPT-3 could be manipulated to generate fake opinions and conversations, and therefore be used for misinformation and disinformation. It also pointed to the system's propensity to produce unsafe content. System 🤖 GPT-3 Operator: Philosopher AI; OpenAI Developer: OpenAI Country: USA Sector: Multiple Purpose: Generate text Technology: Large language model (LLM); NLP/text analysis; Neural network; Deep learning; Machine learning Issue: Mis/disinformation; Ethics; Safetyce; Black box

Date it happened
2020-10-01
Organisation involved
OpenAI; Philosopher AI
Product, system or model
GPT-3; Philosopher AI
Where this came from
Share this incident
XLinkedInFacebookWhatsAppEmail
Attribution

This incident was imported from AIAAIC and is used under CC BY-SA 4.0. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.

This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.