Beyond the culture of the researchers, I think this fundamental difference has also confounded attempts to build rules and regulate this technology. The policymaker still sees a computer and assumes it contains a computer-object, rather than a protean, statistical, myth-object.