The only confidence I have in OpenAI at this point is that they will be using scummy tricks like that all the time. What have they ever done to earn confidence in the other direction?

They can't even lie and blame it on a programming fuckup because they'd have to say AI driven code is buggy.