AI – Why Black Mirrors are Good Things

What do you see when you look in the mirror? I see my flaws first, that I should have had more sleep and that the bathroom mirror, or the morning sunlight is probably too bright. Okay, I lie, it’s England, the morning sunlight is never too bright, if it makes an appearance at all. Sometimes it is more about who I see. I am not about to win any beauty contests, but I like who I see in the mirror. I can see my father’s full upper lip and my mother’s clean jawline. He is, and she was, deeply kind, even when being brutally honest, deeply loving, and endlessly curious. I like seeing them in my face.

As you may know, I am a criminologist and over the years I have dealt with and seen some dark things. Two years ago, I was preparing to do a doctorate on bias in judicial decision-making. This came about chiefly as a result of a heart-breaking year trying to establish what the International Human Rights were for prostitutes (Spoiler Alert: Sod All). Over that year, and a couple of subsequent years spent in that shadow world; a world that so few of us actually see but which apparently collectively fascinates us. This is attested by the way that people react when I tell them that I am a criminologist and the endless conveyor of polished television dramas with artistically placed (and generally whole) victims and some very interesting ways to creatively splatter blood. The reality however is yes, generally overwhelming, as you would expect, but also (sorry) soul-deadening. Most medics I know have quite morbid senses of humour. But I dodged that side of things, despite some juicy offers (sorry) and wound up in the criminal justice system side of things thinking I could make a difference and, after some years of sitting in courts listening to rape and abuse cases, and a whole other type of soul-deadening, it came to me that the outrageously biased decisions that I was witnessing might not be the product of monsters.

Full disclosure, I do have an inside track on this one. Having been attacked in my late twenties whilst on a morning stroll, quite soundly beaten, and left bleeding on a side-walk I know what it is like to go into shock and lose control of your body (thankfully not completely if you take my meaning) but not your mind. As a result, I have a very clear view of how your personal sanctity can vanish in a moment. I also fully understood why those women didn’t fight back. It was that their understanding of the polite world was shattered in a second, they couldn’t conceive of what was happening to them, so they just went into shock; big-eyed bunnies in brutality’s headlights. But the judgements were just horrible to listen to. Shivering fifteen-year-old girls were painted as wanton sirens luring young men into their cunning lairs whilst parents who attended – as ever, so many of the youngsters only had one if any parent there – sometimes cried quietly but generally sat stony faced in the gallery. Why? Simply because those pale and generally thin young girls weren’t covered in bruises like those judges, juries and defending lawyers wanted them to be, that they just lay there. So, in the eyes of those who had never been to that place, they had consented. I was speaking to one of the ushers at the Old Bailey one particularly upsetting day (gang rape) and she revealed, that in five years of ushering there, that she had never seen a rape case successfully prosecuted.

So, why do I not think that these judges were monsters? Because I know (and science backs me up) that most of us really believe that we are fair and good people. I also know that we are all products of our history and our upbringing and that, most of the time we aren’t aware of our own assumptions and biases. Yes, even when they might be blindingly obvious to others (particularly those on the receiving end). So, the intention for my PhD study was to hold up a judgment and decision-making mirror, admittedly a dark one (heavily focused on judgment, get it? Okay, not my best.) but definitely an authoritative, scientifically backed one, to these judges.

The more I understand about artificial intelligence, the more I agree that it is our dark mirror. This is not a new idea, but I don’t think that we have considered the full depth of that mirror. Massive historical data-sets might not definitively reveal my father’s full upper lip and my mother’s jawline (yet) but they can certainly identify where we came from. And where we come from definitely impacts the patterns of our inspirations and in-built biases in both thrilling and mortifying detail. But it goes further, AI is our baby yes, and as a result, it is subject to both nature and nurture. We have identified the historical ‘nature’ in the deep cracks in the data sets but those are quite easy to address. Now we need to consider the nurture. The ‘training’ of AI is being done by young programmers and data scientists and we need to consider whether they need ethical training too. I was incredibly pleased to see that this idea is being looked at in another context recently and was very impressed by both the passion and the compassion of those who spoke of their work with parliamentary committees.

Back to the other point though, don’t mistake my intentions, I don’t want to wade in. I am a huge believer in, and backer of, the young. This is particularly because, as a young person, I had a couple of great mentors and was given massive responsibilities and opportunities to be involved in and influence policy-making way above my measly pay-grade. This was however mainly because I was (almost) the only person of my level, and a few levels above, at that time in my Hong Kong organisation who wrote English fluently (the other guy there just wanted a quiet life.. Ha ha.. Not me!). As a result, I absolutely believe that more young people should have this opportunity. What I see though, especially in the legal and financial worlds is young people’s intellectual fire burning out in years of long hours and mind-crushingly banal workloads. The world wars may have left us alone at this time, but young people are clearly still the cannon fodder for our stupid decisions.

AI can and is helping us here, so, let’s make smart decisions here. How best to use all this burning intellectual property? How would I use it for the judges I described? Well, I am pleased you asked… It might surprise you to know that I actually wouldn’t throw judges out. There are centuries of useful stuff under those manky old wigs which must be tapped and folded into the consciousness of AI. All cases are also not created equal, but yes, some types of cases and judgements could more easily be ‘assisted’. There is just so much potential. There was a fantastic study done by Cornell recently (which is definitely worth a read if heavy methodology and research is your thang) on how prison populations could be substantially reduced with the assistance of machine learning. Many years ago, whilst studying judgment and decision-making I learned of another study in America which was using computers (at that ancient time) to improve the decision-making of young doctors. I believe that the same should be done with anyone in the mix to be a judge. I still think that humans have a role, because we are supposedly more human, yes? And sometimes we are. I think it is great that young data scientists can come into these dusty, mote-filled worlds, bring light and honest reflection, because they need fresh eyes. I still remember the horror, in the first year of my Law studies, that rippled through the huge lecture hall when we discovered that delivering justice bore virtually no relationship to fairness or ethics. But it was more frightening how quickly we came to accept and engage this institutionalised lunacy (definitely a cautionary tale). Personally though it was definitely part of my decision not to practise Law after my graduation. I appreciate the intellectual cut and thrust, but it isn’t really fair if you are never the one who bleeds. Thus, my point about needing uncalcified minds.

I believe we need to give but guide freedom. There is wisdom in our noggins but how best to pass that on without stultifying the result? I am also a believer in mentoring but very few of us are good mentors. I try hard to be and get some wonderful feedback, but it is ever a learning process. Maybe the answer is that we develop a mentoring programme that helps and guides these young minds? Who knows? Bearing in mind that true scientific discovery involves great ideas and a lot of trial and error, we must absolutely guide them to fly, flail and fail. So many questions but all I know for certain is that if we can hope to make the best of this amazing new tool for honesty then we have to look in the mirror and embrace both the good and the evil we see there.