It would be great to see a truly open and truly human benefit focused AI effort, but OpenAI isn't, and as far as I can tell has no chance of becoming, that. Might as well at least try to be an effective company at this point.
Many would disagree.
If you want a for-profit AI enterprise whose conception of ethics is dumping resources into an endless game of whack-a-mole to ensure that your product cannot be used in any embarrassing way by racists on 4chan, then the market is already going to provide you with several options.