AI alignment and safety

Today’s bots can’t turn against us, but they can cause harm. “AI safety” aims to train this tech so it will always