Killer AI, Black Mirrors, & Murphy’s Law

 Debate over the risk and promise of Artificial Intelligence (AI) and  other technologies can be depressingly simplistic, often polarised into  two camps with views that are extreme to the point of caricature. On the  one hand we have those whose understanding of AI seems to have been  predominantly informed by 1980s movies like The Terminator and Wargames,  in which killer AI is bent on the destruction of the human race. At the  other end of the spectrum we have those who seem to believe that AI can  do harm, and that left unregulated it will lead to a risk-free  cornucopia of rainbows and unicorns. There are highly  intelligent people who attempt to explore the middle ground, but they  often have a tendency to get side-tracked by extremely obscure academic  tangents which are of course largely ignored by significant players such  as the military, companies fulfilling lucrative contracts for the  military, and indeed small-medium companies working on applications  which don’t resemble traditional conceptions of AI or robotics. Across  the entire field of AI/technology risk, there is a dangerous myopia at  work. 

Perhaps you have heard of the Netflix series “Black Mirror”, written  by UK media personality, gaming enthusiast and Noir-Futurist Charlie  Brooker. The show is a collection of separate stories, one per episode,  connected only by a common theme of dangers arising from the collision  between society, human nature, and accelerating technological  development. Black Mirror does a remarkable job of addressing the most  subtle – and yet most likely – area of risk from new tech, which is also  the very same area missed by almost all serious researchers of the  issue. “Serious” researchers discuss headline “existential risk” issues  such as the possibility of intelligent industrial processes accidentally  destroying humanity, which is certainly something to think about and  guard against. The subtler kind of problem explored by Black Mirror is  technology’s amplification not only of communication and efficacy, but  of all the human flaws and fragilities which have historically caused  much suffering. 

I’ve put the word “serious” in quotes above because Black Mirror  manages to explore very important ethical, social, and public safety  issues in an entertainment format. Being a fiction-based TV show does  not exclude it from the realm of serious discourse, and public  communication of serious concerns. For example, the show has covered the  ethics of torturing simulated minds (not to mention slavery of those  minds), problems that arise when emotional humans can remember and  perceive things we would naturally forget or miss, and so on. It doesn’t  take much thought to see that these issues can rapidly escalate from  minor nuisances to very serious problems, and that is what Black Mirror  does so well. We just shouldn’t be leaving that conversation entirely to  the realm of popular entertainment (and eventually, politicians with  little or no understanding of the issues). Of course we live in a world  with the beginnings of such problems, already. Our society is governed  according to a democratic ideal, and yet platforms such as Facebook play  an increasing large role in people’s lives while not only controlling  their lives via non-democratic, implicit rules and schemes, but also  while free to change those rules without warning, and increasingly  subjecting such changes to automation and experimentation without  consent. That is a very slippery slope, and we’ve been sliding down it  for some time, now. 

So yes, rogue military AI may be a concern, and industrial processes  run amok are a threat to be guarded against. The bigger, more likely and  more subtle threat, however, is from AI and other tech which doesn’t  try to kill humanity outright but which amplifies our worst tendencies.  Guarding against such a threat will be hard, as it is already with us to  some degree, and “Friendly” AI is not the answer (or at least not the  whole answer). What we need is a general ethical vision for society, or  at least some simple guidelines on what is or isn’t acceptable which can  inform lawmakers. In the absence of properly observed guidelines,  everything that can happen, will happen. We already know that  killer robots are a bad idea, and yet society doesn’t seem particularly  determined to stop the development. How many bigger, less visible  problems await us which we cannot even properly conceive yet, and so are  all the more likely to fully mature in the darkness of our ignorance? 

In this era of rapidly developing high technology, our society can  become anything it is sufficiently determined to be, and that is an  incredible thing. But the inverse is also true, and Murphy’s Law holds:  In this age of apparent impossibilities become commonplace, every possible development we don’t guard against will come to pass, in one way or another. Finally, now, humanity must grow up and take charge of its own boundaries and priorities, as all children eventually must. 

 Written for Metric Media by Mathew Twyman
Co-published on Metric Media's official website