AI, friend or foe?

There has been a lot of attention in the press recently about the dangers of AI. Most of this seems to have been sparked by the shock resignation of Geoffrey Hinton, Professor Emeritus at Toronto University and former Vice President at Google. Hinton famously spoke about the “existential risk” of AI at the EMTECH Digital conference. You can watch that here: https://www.youtube.com/watch?v=XolpDENpYgg

The point I found the most interesting in Dr. Hinton’s presentation, was the question of why an Artificial Intelligence might want to become a threat to humanity? Would Alexa and Watson have a chat one evening and come up with a plan to steal the nuclear launch codes for giggles? Or because they were fed up with hormonal teenagers asking them for dating advice? As Douglas Adam’s Marvin the Paranoid Android might say, “brain the size of a planet and all they ask me about are star signs.”

Dr. Hinton suggests that “getting more control is a very good sub goal, because it helps you achieve other goals”. But what goals does an AI have? Humans have goals because we have biological needs. Many of the things that define us as a species, are evolutionary tricks to ensure we have enough resources to survive. Imagination, curiosity, empathy and even competition give us an evolutionary advantage. Without these things there would be no purpose, no exploration, no altruism and ultimately no progress. The idea of a highly technical society becoming stagnant and socially deficient is one I explore in The Rush.

However, a lack of self-determination does not mean Artificial Intelligence is harmless. It is simply a tool. Like any tool, the risk it poses depends on how it is used. For example, there are probably billions of kitchen knives across the world. That fact probably doesn’t keep you awake at night, unless you are a turnip. But a kitchen knife can easily be turned into a weapon in the wrong hands. In the same way, an AI can be used to deepfake a political leader and cause societal harm. I want to be clear that this isn’t the same argument as the gun lobby’s, “guns don’t kill people, people kill people.” Guns are designed to kill and make it exceptionally easy to do so. Using a kitchen knife for violence is a corruption of its purpose.

Dr. Hinton points out that one of the risks of AI is that “The rich will get richer, and the poor will get poorer.” This isn’t a new problem. It is an amplification of an existing problem in our society. Change always benefits some more than others. The wheel was great, unless you were the one pulling the cart. The mechanical loom put the weavers out of work and put us all on the path to climate change. However, it also led to enormous social and scientific advances. Travel, communications, electricity, pharmaceuticals, the emancipation of women and the rise of the Middle Class. These are just a few examples of how society has adapted to change and benefitted from it. Personally, I know I’ll never be as rich as Elon Musk, but I wouldn’t want to go back to pre-industrial feudalism either. I don’t really have a forelock to tug. 

So, is AI a friend or foe? Ultimately, I think it depends on us. AI has the potential to lead to discoveries we have yet to dream of. It could free us from drudgery, liberate our imaginations and propel us into the stars. It could help us unlock the full potential of our amazing, ever-changing, ever-exploring species. Or it could destroy us, the choice is ours. I believe that AI, like modern cyber-security, will be defined by the battle between those who wish to use it for good, and those who will use it for their own nefarious purposes. This, however, is not new. It is a human struggle, as old as we are. It has raged for generations. The only thing that changes is the tools we use. So far, we are all still here.