They certainly don't have the moral high ground, but in any other context this is usually considered an inference attack--sampling a model somebody else spent a

They certainly don't have the moral high ground, but in any other context this is usually considered an inference attack--sampling a model somebody else spent a lot of money to train in order to build a similar one at a much reduced training cost. So while OpenAI is absolutely lacks the moral high-ground, ByteDance still seems to be engaging in adversarial behavior.