Something interesting to think about: how does accountability affect accuracy threshold? People often feel more comfortable with humans making mistakes that they can blame on them, than AI making mistakes where blame is fuzzy. I think that's the big barrier to self-driving cars (or any other automation where the consequences involve loss of life, e.g. autonomous drones targeting humans; cf "Kill Decision" by Daniel Suarez)
Something interesting to think about: how does accountability affect accuracy threshold? People often feel more comfortable with humans making mistakes that they can blame on them, than AI making mistakes where blame is fuzzy. I think that's the big barrier to self-driving cars (or any other automation where the consequences involve loss of life, e.g. autonomous drones targeting humans; cf "Kill Decision" by Daniel Suarez)