Would OpenAI’s products pass the sniff test as to their legality if they neither make models public nor pass to training data suppliers profits from operating s

Would OpenAI’s products pass the sniff test as to their legality if they neither make models public nor pass to training data suppliers profits from operating said models? There’s still intellectual property and copyright, is there not?