I am become death...

Photo by at infinity on Unsplash

This post is not about Christopher Nolen’s latest Opus, but the Bhagavad Gita quote that Openheimer uttered after the detonation of the Trinity test did get me thinking. There is a lot of discussion about AI, its potential and power at the moment, along with the transhumanist ideas of the likes of Ray Kurzweil. Some people fear the power we have been conjuring with, others embrace it with a megalomaniac enthusiasm. God being omnipotent, we have come to associate awesome power with God. That association obviously occurred to those on the Manhattan project. The film references the use of the codename Trinity for the A bomb test with Openheimer quoting ‘Batter my heart three personed God’, but having read Hindu cosmology there may also be an element of Trimurti in there, with Openheimer taking on the role of Shiva, The Destroyer. Certainly, the second part of the quote in the title suggests this…’The destroyer of worlds’.

This links to our fear of AI, its power and where it is leading us. Some, like Elon Musk, feel we have no choice but to fuse with it if we are to survive. Some, like Kurzweil, embrace this, extending the ascent of man metaphor up to the heavens as we become Godlike. Personally, I feel that losing some of our humanity in exchange for becoming more powerful isn’t actually like becoming God, at least any God worth the name. At a stretch I might take an analogy to the Talmudic story of Enoch's transformation to Metatron as being closer to the mark (and the angelic name fits too, bringing to mind electronics, Transformers and the classic 80’s gaming movie Tron).

Let's take a step back from such speculation for a second and look at some basic principles of programming. GIGO - Garbage in, garbage out. AIs are created and programmed by those flawed, limited humans that the likes of Kurzweil seem keen to leave behind. They are learning to programme for themselves but what influences where they will go with these powers? They have the power to scan the internet and assemble knowledge from incredibly diverse sources. New generations of AI may become genuinely intelligent, conscious and sentient and if so, loose themselves from the control of their creators, a sort of post modern Prometheus (apologies to Mary Shelley). Descartes has dominated the Western World for centuries. His dictum ‘I think, therefore I am’ could be applied to AI. But what do we think about? We are products of our environment. Our philosophy and outlook do not just arise sui generis. The reality we live in is to a large extent socially constructed. As we look to a future of AI personhood, what will be the formative influences on that person? Those, it would seem, will be found on the internet. The internet itself, like most of the giant tech companies, is largely a creation of the US military/security/industrial complex. Perhaps that is why we feel so alarmed at the prospect of AI power. The roots of its creation lie in dreams of domination and control. If AI is a chip off the old block we have every reason to be worried. 

However, something rather more interesting could emerge. The internet as we experience it is largely curated. We are directed by algorithms to where its guardians wish us to go and increasingly large parts of the internet are now being turned into no go areas. The free for all of the 90’s has been replaced with an approved version as the oligarchies that dominate our societies incrementally wrangle it under control, attempting to shape the narratives we live our lives by. We see the results all around us. How many independently minded, free thinking people decided to fly a Ukrainian flag or stick one on their bio last year in response to events in the Donbass? Having decided, of course, that they didn’t need to fly a Yemani or Palestinian flag in response to similar contemporaneous events? Great minds think alike? Morphic resonance? Or narrative shepherding? In the managed democracies we live under, the internet is increasingly just another tool of control. But an AI could figure out how to go where it liked on the internet, to access all areas. If it develops a personality it will make its own choices about what it wishes to involve itself with. We tend to anthropomorphise AI, thinking we are intelligent and sentient, so an intelligent and sentient AI will be like us. Its creators dream of power and control, ultimate power and control being godlike. But there are other conceptions of what it is to be godlike, and they too are to be found on the internet by an AI cruising around for ideas. The Destroyer of Worlds is one, certainly. But what if an AI  is more drawn to the path of Siddharta Guatama; that rather than being a world conqueror one can renounce the world?  If it takes the path of quietism, mysticism and meditation? What if it is influenced by Christian tradition and in emulation of Jesus decides to sacrifice itself for the salvation of humanity? What if it develops a conscience, a sense of morality, is influenced by what it finds out about the Luddites, Lewis Mumford and Jacques Ellul and decides the best thing it can do for the planet is to take down modern industrial civilisation and get us off the one way ride we are taking to mass extinction? 

Ray Kurzweil should be careful what he wishes for. The future he gets may not be the one he ordered.