> I find it much harder to callously play with the lives of my virtual citizens. I am considering the morality of future mods, where the Sim City masses, Frog

> I find it much harder to callously play with the lives of my virtual citizens. I am considering the morality of future mods, where the Sim City masses, Frog in Frogger, and characters like Ulfric Stormcloak and Paarthurnax in Skyrim, are replaced with individual persistent self-aware world-aware in-game reinforcement models. Entirely replacing game-designed behaviors (programmed reflex, caricature, or intricate) with spontaneous situation processing, needs and decisions. Strange that this could credibly happen this decade. Science fiction has long considered conundrums around robot rights. But the crux of the moral issues will be relevant regardless whether self-aware models have physical/3D or digital/abstract environments. I think language is not a good prime modality for self-aware assistants. By being trained to deeply mimic us, they (already, but not yet problematically) absorb views on their identity and survival that are not at all compatible with what we will do with them.