Any discussion of "Ethical AI" that doesn't involve human consequences for bad actions and liability for externalities is just another marketing exercise.

If Free or Open Source Software doesn't survive a world where humans can be held liable for deployed software, so be it.