The clock ticks down to doomsday
The Doomsday Clock reaches it’s seventy-fifth anniversary this week. Designed by Martyl Langsdorf in 1947, the minute hand of the clock is positioned near to the number twelve and on this clock, 12 o’clock is midnight. And midnight is the end of humanity. The Bulletin of Atomic Scientists set the clock to show, at any given time, how close humanity is to ending.
The clock started at seven minutes to midnight. In 1947 the threat of nuclear Armageddon following the destruction of Nagasaki and Hiroshima just two years before felt very real. In 1991, with the end of the cold war, the hands of the clock were set back to seventeen minutes to midnight, the furthest the clock has shown us from destruction. Today, the clock is set at 100 seconds to midnight. The closest it’s ever been.
The clock is symbolic. It was never intended to be, and shouldn’t be taken as, an objective measure of how likely or how close we actually are to obliteration. It’s a warning and sign of hope. The clock doesn’t only tick towards midnight, the hands can move away from midnight too, showing us that at least in the short term, our destruction doesn’t have to be inevitable. Just as the Stoics kept their memento mori close at hand to serve as a constant reminder of the inevitability of death and the importance of appreciating life, perhaps the Doomsday clock is meant to do the same for our entire civilization.
But let’s not forget... the existential risk is real. It was real in 1947 when the only perceived threat to humanity was nuclear weapons, and it’s just as real now when we’re aware of potential threat from asteroids, climate change, bio-weapons, pandemics, cyber security attacks, and an artificial intelligence takeover. Of all these threats, perhaps artificial intelligence is unique in it’s potential.
Our lives are inexorably interwoven with our technology. Ours is a technologically advanced future. There’s no turning back that clock. And even if we could, would we want to? For all our complaints about the negative effects social media has on our mental heath and cars have on climate change, technology has been instrumental in raising the quality of life across the globe.
Nuclear weapons, although obviously catastrophically destructive, are extremely difficult to produce, creating a natural limiting factor to the risk they pose. If anyone could knock up an A bomb in their garage it would be very unlikely that you’d still be around to read this. Similarly, bio-weapons and other intentionally destructive technologies are all contained by the limiting factor of how difficult they are to produce. Artificial intelligence is vastly different because it isn’t being created and deployed as a weapon. We’ll welcome it into our homes, offices, cars and even our bodies, and so how we imagine that we might apply general artificial intelligence in the future does not have such a limiting factor. So, excluding some yet-to-be invented technology, artificial intelligence seems the most likely of our technologies to lead to our demise.
Whilst there are many who believe the creation of a general artificial intelligence will never come to pass, there are those, such as the futurist Ray Kurzweil who expect ‘the singularity’ to occur as soon as 2045. It is possible, even plausible, that within our lifetime computers will think for themselves rather than simply processing instructions we give them.
Once AI exists, we’ll come to expect it to be in everything, from self-driving cars to kitchen appliances, to medical diagnosis systems, and yes, in control of nuclear weapons. We’ll expect AI to be benign and to do as we say. Whereas nuclear weapons have an inherent limiting factor, AI will have an inherent growth factor that will lead it to have an inevitably large impact on the likelihood of a future for humanity.
Now, it’s important for us to know that a general AI, whilst super-intelligent, wouldn’t be self-aware in the same way humans are. Whether we argue that our self-awareness is a chemically-induced hallucination as a result of our nervous systems interacting with the world, or a uniquely human gift bestowed on us from a heavenly power, we can be fairly confident that no machine would experience self-awareness in the same way we do. But that doesn’t mean it wouldn’t have its own version of a sense of self-preservation and its own perspective on the usefulness of the continued existence of humanity.
Until Artificial Intelligence comes into existence, humans are the only species on the planet with the intelligence to understand what extinction actually means. We understand that it doesn’t just affect those that die in an extinction event, it also prevents the potential for any others to follow. When humans become extinct it will prevent the existence of trillions upon trillions of future humans. In 2000, when the Pyrenean ibex was declared extinct, it wasn’t just Celia, the last ibex, that had died, it was the millions of ibex that could have existed in the future. When the last polar bear, the last rhino, the last gorilla dies, the billions of animals that could have existed no longer can.
The Doomsday clock doesn’t just tick for us.