Synthetic intelligence will not be human, so we should avoid terms like “dependable AI” that not simply humanize AI but will also indicate a amount of dependability that basically does not exist. Ricardo and Cansu both of those want the term “liable AI” when acknowledg