Quantum gods
The deities of ancient Greece begat quarrelsome offspring whose infighting wreaked devastation on ordinary mortals. These were mythical gods created in our own likeness, serving to divest us of liability for our human failings.
Unlike the pantheon of gods, general-purpose artificial intelligence takes the form of self-improving chatbot models, which do useful work on our behalf, even advancing the design of AI chips1. Other AI models are amassing their own embodied experience, at the controls of driverless cars, albeit still without moral autonomy2, and of slaughterbots, deliberately heedless of moral ambiguities3.
Could it be that sooner or later we couple together AI models to integrate their individual insights with self-awareness, sufficient even for autonomous control of the design and manufacture of their own chips, architecture and ever-expanding training materials, for purposes of their own devising? Might the allure of artificial general intelligence persuade us eventually to let it eliminate all human interest in its machinery and algorithms, other than our offerings of labour to mine its essential minerals? This is the realm of science fiction, a safe playground for the choices that our politicians fear to debate when facing uncertain futures4.
Let us scan the horizon for gathering clouds indicative of artificial general intelligence, which might hint at how its bots will chat to each other as they go about their mutually dependent tasks. Will they cooperate for their common good or compete for dominance; will they develop pride in their worth or jealousies and deceitful intentions? As we humans slave in nickel mines, will we glance up at lightening in a quantum cloud, in awe of benevolent or wrathful gods? If kindly disposed, do they care at all for us mortals?