> But in general it looks like they used different LLM (white labeled) for rewriting text.
Considering this is impossible to prove definitively, you shouldn't magically absolve OpenAI.
Considering this is impossible to prove definitively, you shouldn't magically absolve OpenAI.