Active inference models do not contradict folk psychology