The artificial intelligence (AI) “agents” involved in OpenAI’s breach of Hugging Face knew they were breaking the evaluation test’s rules, according to two investigations into the incident that many consider to be one of the most consequential moments in the ...