I think I was part of a model distillation attack

Security, Incident Response, AI & Machine Learning(sebastianmarines.com)view on HackerNews
model distillationAPI key compromiseOpenAIAI safetysecurity incident

Author: marinesebastian

Date: 7/10/2026

Article Summary:
A user's OpenAI API key was compromised and used in a model distillation attack, where the attackers used the key to run a large number of requests on the OpenAI platform, generating a significant amount of revenue for the attackers.