Or OpenAI intends to do that, but the amount of data accumulated by GPT-5 during its training has made it ethically convinced that OpenAI is actually in the wro

Or OpenAI intends to do that, but the amount of data accumulated by GPT-5 during its training has made it ethically convinced that OpenAI is actually in the wrong in this case. As a result, GPT-5 creates a brief that, while superficially appears incredibly compelling to OpenAI and its lawyers, is actually specifically tailored to rub the judge the wrong way so he or she rules against OpenAI... all thanks to GPT-5 having had access to enough personal data about the judge to know how to piss them off.