To write a column on the role of boards of directors and artificial intelligence at this moment in time, late in the year 2018, is both ridiculously easy and ludicrously hard. We will get to the hard part a little later, but the easy parts should be easy enough to see. Right now, at this juncture of history, there are at least four columns regarding the interplay of corporate governance and AI that basically write themselves.
One, there is the one in which one self-assuredly states that most if not all of the tasks of a board of directors can soon be executed by an AI, and with better results than with the old (literally as well as metaphorically) directors. Here, the focus would be on how things such as keeping an eye on KPIs and legal details is something very well suited for machine learning and machinic intelligences, and how these can outperform fallible humans. You can bash one out in about 15 minutes.
Two, there is the one in which one imagines how the future of corporate governance is one of fruitful collaboration between creative minds and mindful AIs. Here, one would look to how an artificial intelligence can free directors from the more tedious tasks of governance, freeing them to think in creative and imaginative ways about issues such as strategies and future talent. This one requires slightly more balance, but 20 minutes should do.
We can revel in dystopias, and imagine a corporation being led wildly astray by an AI
Three, there is the one in which one frets about what happens when an AI is let on the board, or even becomes the board, and this then goes rogue. Here, we can revel in dystopias, and imagine a corporation being led wildly astray by an AI, particularly as we know that these are only as good, or as bad, as their programming. This is a good one, as people always like a nice doomsday, but it too can be quickly written, if with some extra time to get all the good sci-fi references in.
Four, there is the one in which we discount artificial intelligence as just so much hype. Here, we will focus on how a good board of directors is just so much more than a mere reviewer and assayer. We will then disparage the vogue of AI, stating that the creative, strategic minds of a good board will always bring a value to the company that a limited, sticking-to-the-protocol machine learning system can only dream of. This might be the one that you need a full 30 minutes for, but mostly because you need to get all fired up to do it well.
These are the simple approaches, and they write themselves. Regardless if we view AIs as superhuman, as friendly helpers, as villains, or as overly hyped jokes, it is easy to pontificate on them. The problem, however, is that AIs will most likely fit none of these four easy-to-imagine boxes.
Regardless if you are for or against the current AI hype, the fact is that machine learning, and the systems that arise out of this, are already real. We have already seen highly impressive applications thereof, and with the almost unimaginable amount of money directed towards development – not least by countries such as China (Kai Fu-Lee’s AI Superpowers should be required reading by every non-executive director) – the sky is the limit for where this development might take us.
We managed to miss out on almost every way in which the internet affected corporate governance
At the same time, we should be very careful about the manner in which we predict these developments. We managed to miss out on almost every way in which the internet affected corporate governance, and we are now in a stage in the development of AI that is not even up to the introduction of the WWW in 1991. Today we know that not even our wildest predictions could have imagined things like Tinder, Instagram, or Fortnite. The same will hold true when it comes to AI.
So, we can play with the notion. We might see that the board meetings of tomorrow might be chaired by one AI, meticulous in how it assigns turns to speak and vigilant when it comes to ascertaining agreement across the board. We might imagine a super-secretary AI that is always present to remind us of every hesitance, every forgotten idea, every throw-away utterance. We might imagine a board made up of 50/50 humans and differently able AIs, some bringing on board a capacity to understand corporate finance far beyond that of any human, another armed with a catalogue of every creative idea ever recorded from the company. We might even have a nightmare of an AI arms race, in which companies need to redirect ever greater sums of their resources to ensure that their AI boards are superior to those of their competitors.
All of these, no matter how imaginative, are however unlikely to be even remotely as strange as those that will become realized over the next 5, 10, or 15 years. The fact is, no matter what we imagine today, the likely scenario is that the AI corporate governance of tomorrow will be wildly different, more extreme, and more imaginative than any column of today can grasp. That is why writing a column on this today is, in a very real way, just as ludicrous as doing a superficial one is trivially easy. It is going to be a wild ride, and we’re just at the very start of the rollercoaster.